BIDIMENSIONAL EMPIRICAL MODE DECOMPOSITION USING VARIOUS INTERPOLATION TECHNIQUES

Size: px
Start display at page:

Download "BIDIMENSIONAL EMPIRICAL MODE DECOMPOSITION USING VARIOUS INTERPOLATION TECHNIQUES"

Transcription

1 Advances in Adaptive Data Analysis Vol. 1, No. 2 (2009) c World Scientific Publishing Company BIDIMENSIONAL EMPIRICAL MODE DECOMPOSITION USING VARIOUS INTERPOLATION TECHNIQUES SHARIF M. A. BHUIYAN, NII O. ATTOH-OKINE, KENNETH E. BARNER, ALBERT Y. AYENU-PRAH and REZA R. ADHAMI, Department of Electrical and Computer Engineering University of Alabama in Huntsville Huntsville, AL 35899, USA, Department of Civil and Environmental Engineering University of Delaware Newark, DE 19716, USA Department of Electrical and Computer Engineering University of Delaware Newark, DE 19716, USA bhuiyas@ece.uah.edu okine@ce.udel.edu barner@ece.udel.edu albert@udel.edu adhami@ece.uah.edu Scattered data interpolation is an essential part of bidimensional empirical mode decomposition (BEMD) of an image. In the decomposition process, local maxima and minima of the image are extracted at each iteration and then interpolated to form the upper and the lower envelopes, respectively. The number of two-dimensional intrinsic mode functions resulting from the decomposition and their properties are highly dependent on the method of interpolation. Though a few methods of interpolation have been tested and/or applied to the BEMD process, many others remain to be tested. This paper evaluates the performance of some of the widely used surface interpolation techniques to identify one or more good choices of such methods for envelope estimation in BEMD. The interpolation techniques studied in this paper include various radial basis function interpolators and Delaunay triangulation based interpolators. The analysis is done first using a synthetic texture image and then using two different real texture images. Simulations are made to focus mainly on the effect of interpolation methods by providing less or negligible control on the other parameters or factors of the BEMD process. Keywords: Bidimensional empirical mode decomposition; intrinsic mode function; scattered data interpolation; radial basis function; Delaunay triangulation; nonlinear and non-stationary data analysis. 1. Introduction The empirical mode decomposition (EMD) technique has been developed with a view to analyze time-frequency distribution of nonlinear and nonstationary data. 1,2 309

2 310 S. M. A. Bhuiyan et al. It is an adaptive decomposition with which any complicated signal can be decomposed into its intrinsic mode functions (IMFs), providing well-defined instantaneous and/or local frequency information about a signal. This decomposition technique has also been extended for analyzing images or 2D data, which is known as bidimensional EMD (BEMD), image EMD (IEMD) 2D EMD, etc BEMD is a promising image processing algorithm that can be applied in various real-world problems, e.g., medical image analysis, pattern analysis, and texture analysis, Both EMD and BEMD require finding local maxima and minima points and subsequent interpolation of those points at each iteration of the process. One dimensional (1D) extrema points are obtained using either a sliding window or a local derivative; 2D extrema points are obtained using a sliding window or various morphologic operations. 1 5 Cubic spline interpolation is preferred for 1D interpolation, while thin-plate spline, B-spline, Delaunay triangulation with cubic interpolation, finiteelement method, etc., have been used as 2D scattered data interpolation for envelope surface estimation in BEMD, 1 7 where Delaunay triangulation and finite-element method based surface estimation result in relatively faster decomposition than the others. Beside true 2D implementation of the BEMD process, 1D implementation of the BEMD process is also studied in the literature, where each row and/or each column of the 2D data is actually processed by 1D EMD, which makes it a faster process However, it has been found that this 1D implementation results in poorer IMF components compared to the 2D implementation due to the fact that the former ignores the correlation among the rows and/or columns of a 2D image. 11 BEMD decomposition and the resulting 2D IMFs are governed by the method of extrema detection, criteria for stopping the iterations for each IMF and interpolation techniques. Though all of these factors are important for successful decomposition, the interpolation method may be considered the most crucial. Most of the scattered data interpolation techniques to produce 2D surfaces are themselves iterative processes. In the case of BEMD, it is very likely that the maxima or minima map does not contain any interpolation centers at the boundary region, which may be more severe for the later modes of decomposition. Hence, some kind of boundary processing to introduce additional interpolation centers at the boundary may also be required for successful decomposition. 11,12 Interpolation of the local maxima points is needed to form the upper envelope, and interpolation of the local minima points is needed to form the lower envelope of the data/image. The average of the upper and the lower envelopes of the image gives the mean envelope. One of the purposes of the BEMD decomposition is to get 2D IMFs with zero local mean defined by the mean envelope, which further plays a significant role for orthogonal decomposition. Hence, the accuracy of the envelopes in terms of shape and smoothness is very important, which calls for the need to identify an appropriate 2D scattered data interpolation technique for BEMD.

3 Bidimensional Empirical Mode Decomposition 311 There are a few 2D scattered data interpolation techniques that are employed for the BEMD process so far, and there are few works that compares the BEMD performance for few interpolation techniques. 13,14 But there is no work in the literature that compares BEMD performance for many interpolation techniques with detailed simulation/analyses. In this paper, BEMD simulation is performed using radial basis function (RBF) based interpolations and Delaunay triangulation (DT) based interpolations. To determine the true effect of interpolation techniques, boundary processing is not employed in this analysis. In fact, boundary processing is computationally extensive, and there may not be a universally standard technique to accomplish this. Hence, an interpolation technique that results in a good decomposition even without boundary processing of the data may be preferable. In this paper, a synthetic texture image with three component textures has been used for the initial simulation. Then, two real texture images are used for further verification of the results. Various metrics are considered to evaluate the effectiveness or performance, or both, of each of the interpolation techniques for the case of the BEMD process. The results may be used as a guideline for selecting an appropriate interpolation for BEMD based on specific goals and system constraints. The rest of the paper is organized as follows. Section 2 gives a brief overview of the BEMD process. Section 3 outlines the interpolation techniques used in this study. Simulation results are given in Sec. 4, while analysis and discussion of the results are given in Sec. 5. Finally, concluding remarks are given in Sec BEMD Overview EMD as well as BEMD decomposes a signal into its IMFs based basically on the local frequency or oscillation information. The first IMF contains the highest local frequencies of oscillation, the final IMF contains the lowest local frequencies of oscillation and the residue contains the trend of the data. Application of Hilbert transform to the 1D IMFs produces analytic signals that are used to analyze the time frequency distribution known as the Hilbert spectral analysis (HSA); EMD plus HSA together form what is known as the Hilbert Huang transform (HHT). 1 It is claimed and experimentally shown that the HHT performs better than the other existing techniques of analyzing the time frequency distribution of nonstationary and nonlinear data. 1 Like time frequency distribution with EMD, acquiring the space spatial frequency distribution of 2D data/image may be possible with BEMD. Although direct estimation of the horizontal and vertical frequencies of 2D IMFs has been studied, 15 2D HHT has not yet been reported in the literature. The IMFs obtained from EMD are expected to have the following properties 1 5 : (i) in the whole data set, the number of extrema and the number of zero crossings must be equal or differ by at most one; (ii) there should be only one mode of oscillation between two successive zero crossings; (iii) at any point, the mean value of the envelope defined by the local maxima and the envelope defined by the local

4 312 S. M. A. Bhuiyan et al. minima is zero; and (iv) the IMFs are locally orthogonal. In fact property (i) ensures property (ii), and vice versa. The definition and properties of the BIMFs are slightly different from the IMFs. It is sufficient for BIMFs to follow only the final two (iii and iv) properties given above. 3,4 In fact, due to the properties of an image and the BEMD process, it is not possible to satisfy the first two properties (i and ii) given above in the case of BIMFs, since the maxima and minima points are defined in a 2D scenario for an image. Let the original image be denoted as I, a2dimfasf, and the final residue as R. In the process ith IMF F i is obtained from its source image S i (alternately known in the literature as intermediate residue), where S i = S i 1 F i 1 and S 1 = I. It requires one or more iterations to obtain F i, where the temporary state of an IMF (TS-IMF) in jth iteration can be denoted as F Tj. Using the above-mentioned definitions, the steps of the BEMD process can be summarized as below 1 5 : (i) Set i =1andS i = I. (ii) Set j =1andF Tj = S i. (iii) Obtain the local maxima and minima maps of F Tj, denoted as P j and Q j, respectively. (iv) Form the upper envelope (UE) and lower envelope (LE) of F Tj, denoted as U Ej and L Ej, from the maxima and minima points in P j and Q j, respectively. (v) Find the mean envelope (ME) as M Ej =(U Ej L Ej )/2. (vi) Calculate F Tj+1 as F Tj+1 = F Tj M Ej. (vii) Check if F Tj+1 follows the 2D IMF properties. This can be ensured by finding the standard deviation (SD), denoted as D, between F Tj+1 and F Tj. 1 5,16,17 D may be defined in the following different ways 1 5 : D = D = M x=1 y=1 M x=1 y=1 N F Tj+1 (x, y) F Tj (x, y) 2 F Tj (x, y) 2, (1a) N F Tj+1 (x, y) F Tj (x, y) 2 F Tj+1 (x, y) 2, (1b) D = M x=1 N y=1 F Tj+1(x, y) F Tj (x, y) 2 M x=1 N y=1 F Tj(x, y) 2, (1c) where (x, y) is the coordinate and M (N) is the total number of rows (columns) in the 2D data. Although all three SD measures in Eqs. (1a), (1b), and (1c) provide a global measure of SD, the third one causes the fastest convergence while the second one causes the slowest convergence. It should be noted that the definition of SD affects the number of required iterations to achieve a given threshold and thus the amount of sifting per IMF; it does not have any contribution to the calculation of UE, LE, or ME in a particular iteration. To satisfy the nearly zero local envelope mean for each IMF, the

5 Bidimensional Empirical Mode Decomposition 313 SD threshold can be set below 0.5 for Eq. (1b), and below 0.05 for Eqs. (1a) and (1c) for practical purposes. (viii) If F Tj+1 meetsthecriteriagiveninstep(vii),thentakef i = F Tj+1 ;set S i+1 = S i F i and i = i+1; go to step (ix). Otherwise, set j = j+1, go to step (iii) and continue up to step (viii). (ix) If S i has less than three extrema points then the final residue R = S i ;and the decomposition is complete. Otherwise, go to step (ii) and continue up to step (ix). Except for the truncation error of the digital computer, the summation of all the IMFs and the final residue, denoted as C, returns the original data/image back as given by K C = F i + R = I, (2) i=1 where K is the total number of IMFs excluding the residue. Following Ref. 1 an orthogonality index (OI), denoted as O, may be defined for 2D IMFs and residue as follows: O = M x=1 y=1 N K+1 i=1 K+1 j=1 F i (x, y)f j (x, y) 2, (3) C (x, y) where F K+1 denotes the final residue R. A low value of OI indicates a good decomposition in terms of local orthogonality among the components. The whole process of decomposition of an image into its 2D IMFs in order of local spatial scales is known as sifting. The decomposition of any image into 2D IMFs is not a unique process. The number of IMFs and their characteristics depends on the extrema detection method, interpolation technique, and stopping criteria of the iterations for each IMF. In that sense, there are an infinite number of IMF sets for each image. 16 Although the SD threshold criterion is used to satisfy the zero envelope mean for each IMF, this criterion sometimes causes over sifting of the data and results in a poor decomposition 16,17 causing the IMF to be a constant-amplitude frequency modulated function lacking the intrinsic amplitude variations. To overcome this problem, various additional stopping criteria may be employed. 16,17 3. Interpolation Techniques Surface interpolation from 2D scattered data (local maxima and minima points) is a crucial and important part of BEMD method. Scattered data interpolation refers to the computation of a suitable surface model that approximates arbitrarily distributed discrete data samples. The process of reconstructing smooth surfaces from discrete data can be achieved either by interpolation or by approximation.

6 314 S. M. A. Bhuiyan et al. Since the solution of the interpolation problem coincides with the data points, one might assume it to be the more accurate reconstruction. But if the samples carry some noise, an approximating surface will be more adequate. Therefore, the choice of the appropriate method depends on the specific structure of the problem In this section several RBF based surface interpolation and several DT based surface interpolation techniques are reviewed in brief. Some of these interpolation methods are later tested in terms of BEMD performance in this paper RBF based interpolation RBF based interpolation methods are examples of global interpolation methods for scattered data points. RBF methods impose fewer restrictions on the geometry of the interpolation centers and are suited to problems where the interpolation centers do not form a regular grid as in the case of local maxima or minima maps of images or textures. In addition, RBF methods are some of the most elegant schemes from a mathematic point of view. 18,21 RBFs are well known to provide powerful tools for high-fidelity reconstruction of surfaces from a selected set of sparse and irregular samples, in which context they have intensively been studied for more than twenty years. 22,23 However, a maximum number of interpolation centers with which an RBF interpolator can work may pose some limitations on its usefulness. Classical examples for these functions are given by Duchon s thin plate splines 24,25 and Hardy s multiquadrics, 21 where developments during this decade have provided compactly supported RBF (CSRBF) and RBFs with multizone decomposition. Consider f : R d R a real-valued function of d variables that is to be approximated by s : R d R, given the values {f(x i ) : i = 1, 2, 3,...,N}, where {x i : i =1, 2, 3,...,N} is a set of distinct points in R d called the nodes of operation. The RBF approximation s is given by where s(x) =p m (x)+ N λ i φ( x x i ), x R d, λ i R, (4) i=1 s is the radial basis function (RBF); p m is a low-degree polynomial, typically linear or quadratic, a member of the mth degree polynomials in d variables; denotes the Euclidean norm; the λ i s are the RBF coefficients; φ is a real-valued function called the basis function; and the x i s are the RBF centers. Some examples of popular choices of φ and the names of the corresponding RBFs are given in Table 1. In Table 1 the constant c for the cases of Hardy s multiquadrics, inverse multiquadrics, exponential splines and Gaussian splines may be selected based on specific need and performance. 26

7 Bidimensional Empirical Mode Decomposition 315 Table 1. Choices of φ for various RBFs. RBFs Choices of φ Linear φ(r) =r Cubic splines φ(r) =r 3 Thin-plate splines φ(r) =r 2 log(r) Hardy s multiquadrics φ(r) = r 2 + c 2a Inverse multiquadrics φ(r) = 1 r 2 +c 2 Exponential splines φ(r) =e cr Gaussian splines φ(r) =e cr2 Compactly supported splines φ(r) =(1 r) mb a c is a constant that governs the shape/spread of the basis function b m is a function of spatial dimension 3.2. DT based interpolation Given a set of data points, the DT is a set of lines connecting each point to its natural neighbors. The DT is related to the Voronoi diagram the circle circumscribed about a DT has its center at the vertex of a Voronoi polygon. Voronoi diagram refers to the partitioning of a plane with n points into convex polygons such that each polygon contains exactly one generating point, and every point in a given polygon is closer to its generating point than to any other. 27 A Voronoi diagram is sometimes also known as a Dirichlet tessellation. The cells are called Dirichlet regions, Thiessen polytopes, or Voronoi polygons. The DT is a triangulation that is equivalent to the nerve of the cells in a Voronoi diagram, i.e., that triangulation of the convex hull of the points in the diagram in which every circumcircle of a triangle is an empty circle. 28 The DT can be used to create a triangular grid for scattered data points. Subsequent piecewise interpolation on triangles can be used to create an approximated surface. 6,29 The interpolation methods for this purpose may be linear, cubic or nearest neighbor. 4. Simulation Results RBF based and DT based 2D interpolation methods are suitable for the type of scattered data appearing in the maxima or minima map in the BEMD process Hence, these two types of interpolation techniques are considered in this paper. A comprehensive study of BEMD process has been made for these two types of major interpolation techniques using synthetic as well as real texture images. Extrema (maxima and minima points) detection is required prior to the interpolation. In this simulation extrema points are found by using neighboring window method employinga3 3 window. Generally, a 3 3 window results in an optimum extrema map for a given 2D data. In this process a 3 3 window is moved all over the image/2d data points. If the pixel corresponding to the center of the window has a value that is strictly higher than all of its neighbors (i.e., eight neighboring pixels) within the window, then it is considered a local maximum; similarly, if the same

8 316 S. M. A. Bhuiyan et al. pixel has a value that is lower than all its neighbors, then it is considered a local minimum. In this work, a slightly different neighboring window is considered for the corner and boundary pixels rather than ignoring them for finding local maxima or minima, where each corner pixel has only three neighbors and each boundary pixel has only five neighbors to be considered. This method may result in one or more maxima/minima points at the boundary region without any other heuristic processing and thus may reduce the boundary effects of the interpolation in the BEMD process. In the local maxima (minima) image, all but the local maxima (minima) points are set to zero. An SD threshold is the fundamental criterion for stopping the iterations for a particular IMF. To avoid over-sifting, various additional stopping criteria may be employed as for the case of 1D EMD. 16,17,30 In this simulation, besides the original SD threshold criterion, a maximum number of iterations for each IMF is set as an additional criterion. However, this additional criterion is set in such a way that the decomposition is mostly governed by the interpolation technique, not by the stopping criteria BEMD simulation using synthetic texture Synthetic texture The synthetic texture is composed of three components. Each component is synthesized from horizontal and vertical approximate sinusoidal waveforms having different but closely spaced frequencies. The first texture component consists of higher frequencies, the second component consists of medium frequencies, and the last component consists of very low frequencies. The amplitudes and frequencies of various components are given in Table 2. The OI among the original synthetic texture components and the mean value of each component are given in Table 3. These values are also useful for comparing the resulting IMFs with the actual components. For the simulation, the synthetic texture is chosen to be pixels. Hence, the sampling frequency, or the number of data points, is taken accordingly. Onedimensional horizontal and vertical sinusoidal components are shown in Figs. 1 and 2, respectively. 2D plots of each texture component along with their horizontal and vertical elements are shown in Figs The final composite synthetic texture image is given in Fig. 6. Diagonal 1D slices of all three texture components are Table 2. Amplitude and frequencies of 1D sinusoidal waves used to generate the synthetic texture. Horizontal Vertical Component 1 Amplitude Frequency Component 2 Amplitude Frequency Component 3 Amplitude Frequency

9 Bidimensional Empirical Mode Decomposition 317 Table 3. Global mean of synthetic texture components and their orthogonality index. Global mean Orthogonality index (OI) Component Component Component Fig. 1. 1D sine wave components in the horizontal direction of the synthetic texture. Fig. 2. 1D sine wave components in the vertical direction of the synthetic texture. shown in Fig. 7. Although formation of the synthetic texture components using horizontal and vertical approximate sinusoidal waveforms gives regular texture grids, only the BEMD can directly generate similar texture components back from the composite synthetic texture, whereas the Fourier decomposition fails to do so. For example, the pixel composite synthetic texture image has 256 frequency components in the horizontal direction and another 256 frequency components in (a) (b) (c) Fig. 3. For highest frequency synthetic texture component: (a) 2D texture from horizontal sine waves, (b) 2D texture from vertical sine waves, and (c) 2D texture from addition of (a) and (b).

10 318 S. M. A. Bhuiyan et al. (a) (b) (c) Fig. 4. For medium frequency synthetic texture component: (a) 2D texture from horizontal sine waves, (b) 2D texture from vertical sine waves, and (c) 2D texture from addition of (a) and (b). (a) (b) (c) Fig. 5. For lowest frequency synthetic texture component: (a) 2D texture from horizontal sine waves, (b) 2D texture from vertical sine waves, and (c) 2D texture from addition of (a) and (b). the vertical direction. Thus it can be decomposed into 65,536 component images by applying Fourier decomposition, where each component image corresponds to a combination of one horizontal frequency element and one vertical frequency element. The reconstruction/approximation of the original texture image requires addition of hundreds of dominating image components obtained by Fourier analysis Extrema detection Detection of maxima and minima points is required for each iteration of the process. As the source image changes from one iteration to the next, the extrema points also change in number and sometimes in position at each iteration. However, for a clear understanding of the extrema detection, the minima and maxima images and their 3D mesh plots for the original image are given in Figs. 8 and 9, respectively.

11 Bidimensional Empirical Mode Decomposition 319 Fig. 6. The pixel synthetic texture image Interpolation Scattered data interpolation is used to interpolate the maxima and the minima points to produce 2D continuous surfaces, called the upper and the lower envelopes, respectively. The average of these two envelopes results in the mean envelope, which indicates the local mean of the data. Since the maxima and the minima maps change in each iteration, the corresponding interpolated surfaces should also change at each iteration. However, for illustration purposes, the interpolated envelopes along with the original data for various interpolation methods are given in Fig. 10 for the original synthetic texture image. In Fig. 10, the left column shows 3D mesh plots of a32 32-pixel region of the original data and corresponding envelopes, while the right column shows the diagonal slices (intensity profiles) of the original data and corresponding envelopes. From Fig. 10, it is found that the RBF interpolators produce complete 2D surfaces that might be useful for the BEMD process. On the other hand, most of the DT based interpolators can produce a surface within a rectangle region that bounds the largest x- andy-coordinates of the maxima/minima data. Hence, in its original form, DT cannot be used for BEMD for all kinds of data. In cases where it may be possible to have some maxima or minima points at the extreme coordinates as in Ref. 6, DT may be suitable over the other techniques, still with the addition of some data points to meet the requirement of convexhull. But in the iterative process it is not certain that there will be some extrema points at the boundaries, and it may also be impractical to put in additional data points for this interpolation. Although DT followed by the nearest neighbor interpolation can produce a surface up to the boundary, the interpolation is very poor and fails to produce meaningful IMF components. Based on these facts, DT may not be suitable, in general, for the BEMD process and is not considered for the subsequent analysis of this paper Results For the simulation using synthetic texture, an SD threshold of 0.5 and a maximum of 1000 iterations are used as the stopping criteria. For RBF interpolators where

12 320 S. M. A. Bhuiyan et al. (a) (b) (c) (d) Fig. 7. 1D diagonal slices (intensity profiles) of the synthetic texture image and its components: (a) original texture, (b) component-1, (c) component-2, and (d) component-3. Fig. 8. Minima image and corresponding mesh plot of the original synthetic texture image. needed, the inverse of the approximate average distance among the interpolation centers is used as the value of the constant c. The IMFs and residue obtained from the synthetic texture for RBF interpolation techniques are shown in Fig. 11, where it is observed that some techniques result in more than three components. Figure 12

13 Bidimensional Empirical Mode Decomposition 321 Fig. 9. Maxima image and corresponding mesh plot of the original synthetic texture image. (a) RBF-linear (b) RBF-cubic Fig. 10. Envelopes and original image: left, 3D mesh plots of a pixel region and right, diagonal slices (intensity profiles) of the original image and corresponding envelopes.

14 322 S. M. A. Bhuiyan et al. (c) RBF-thin-plate spline (d) RBF-multiquadric (e) RBF-inverse multiquadric Fig. 10. (Continued )

15 Bidimensional Empirical Mode Decomposition 323 (f) RBF-exponential (g) RBF-Gaussian (h) DT-linear Fig. 10. (Continued )

16 324 S. M. A. Bhuiyan et al. (i) DT-cubic (j) DT-nearest neighbor Fig. 10. (Continued ) depicts the diagonal slice (intensity profile) of each component for the interpolation methods that result in only three components so that they can be compared with the original intensity profiles. The total number of IMFs including the final residue, total number of iterations needed, and OI for various RBF interpolation techniques are summarized in Table 4. To compare the resulting IMFs with the corresponding original components, several metrics are considered, which are peak signal-to-noise ratio (PSNR), mean square error (MSE), mean absolute error (MAE), and correlation coefficient (CC). For the interpolation methods, which generate more than three components, even though the first IMF bears some resemblance to the actual first component of the synthetic texture, the other two original components cannot be compared with more than two remaining IMF components. Consequently, IMF-wise iterations, achieved SD, global mean, MSE, MAE, CC, and PSNR are displayed in Tables 5 11 only for the interpolation techniques that result in three components. As mentioned previously, these parameters are used to compare the IMF components with the actual components. Although this comparison is possible only for the synthetic case, it gives a picture of the performance of BEMD process for various interpolation techniques.

17 Bidimensional Empirical Mode Decomposition 325 (a) RBF-linear (b) RBF-cubic (c) RBF-thin-plate spline Fig D IMFs and final residue obtained from the synthetic texture for different RBF interpolators.

18 326 S. M. A. Bhuiyan et al. (d) RBF-multiquadric (e) RBF-inverse multiquadric (f) RBF-exponential (g) RBF-Gaussian Fig. 11. (Continued )

19 Bidimensional Empirical Mode Decomposition 327 (a) RBF-cubic (b) RBF-thin-plate spline (c) RBF-inverse multiquadric (d) RBF-Gaussian Fig D diagonal slices (intensity profiles) of the 2D IMFs, final residue, and their sum for different RBF interpolators. Table 4. Comparison among different RBF interpolators with respect to the synthetic texture image of Fig. 6: total IMFs, total iterations, and OI. RBF interpolators Total number of IMFs/components Total iterations needed OI Linear Cubic Thin-plate spline Multiquadric Inverse multiquadric Exponential Gaussian BEMD simulation using real images Different interpolation techniques are next applied to several real texture images. In this case, the simulation parameters for various steps of the BEMD process are kept the same as for the case of synthetic texture except the SD threshold which

20 328 S. M. A. Bhuiyan et al. is set to 10. A higher value of the SD threshold is chosen to limit the number of iterations. Since the SD value after the first iteration is very high, the SD threshold of 10 may be considered acceptable. It may not ensure a perfect local zero mean for the IMFs, but it may be better than applying additional stopping criteria in the sense that any additional stopping criteria can cause an even higher stopping point SD value. 16 Hence, the SD threshold value of 10 can adequately serve the purpose of finding the effect of interpolation methods on the BEMD process. For the true images it is not possible to compare the decomposed IMFs with the actual or true components, since there are no ground-truth IMFs. Hence the total number of Table 5. Comparison among different RBF interpolators with respect to the synthetic texture image of Fig. 6: iterations needed. RBFinterpolators IMF1 IMF2 Cubic Thin-plate spline Inverse multiquadric Gaussian Table 6. Comparison among different RBF interpolators with respect to the synthetic texture image of Fig. 6: achieved SD. RBF interpolators IMF 1 IMF 2 Cubic Thin-plate spline Inverse multiquadric Gaussian Table 7. Comparison among different RBF interpolators with respect to the synthetic texture image of Fig. 6: global mean. RBFinterpolators IMF1 IMF2 Residue Cubic Thin-plate spline Inverse multiquadric Gaussian Table 8. Comparison among different RBF interpolators with respect to the synthetic texture image of Fig. 6: MSE. RBF interpolators IMF 1 IMF 2 Residue Sum of all components Cubic e 028 Thin-plate spline e 028 Inverse multiquadric e 028 Gaussian e 028

21 Bidimensional Empirical Mode Decomposition 329 Table 9. Comparison among different RBF interpolators with respect to the synthetic texture image of Fig. 6: MAE. RBF interpolators IMF 1 IMF 2 Residue Sum of all components Cubic e 015 Thin-plate spline e 015 Inverse multiquadric e 015 Gaussian e 015 Table 10. Comparison among different RBF interpolators with respect to the synthetic texture image of Fig. 6: CC. RBF interpolators IMF 1 IMF 2 Residue Sum of all components Cubic Thin-plate spline Inverse multiquadric Gaussian Table 11. Comparison among different RBF interpolators with respect to the synthetic texture image of Fig. 6: PSNR. RBF interpolators IMF 1 IMF 2 Residue Sum of all components Cubic Thin-plate spline Inverse multiquadric Gaussian obtained IMFs/components, the OI, the total number of iterations, and the number of iterations needed for the first five IMFs are considered as the performance indices. The images used are two texture images from the Brodatz texture set, 31 namely D18 and D27. Each of the images used in the simulation is pixels, as shown in Fig. 13; the images are taken as cropped regions from the original pixel images. As sample representation, the generated IMFs for D27 are shown in Figs. 14(a) 14(g) for various RBF interpolators. The comparison among all the (a) (b) Fig. 13. The pixel real texture images taken from Brodatz texture set [28]: (a) D18 and (b) D27.

22 330 S. M. A. Bhuiyan et al. (a) RBF-linear Fig D IMFs and final residue obtained from real texture image D27 for different RBF interpolators.

23 Bidimensional Empirical Mode Decomposition 331 (b) RBF-cubic (c) RBF-thin-plate spline Fig. 14. (Continued ) different RBF interpolators for BEMD of D18 is given in Table 12 and for BEMD of D27 is given in Table Analyses and Discussion For the synthetic texture, each of the RBF-based interpolation techniques results in a different set of IMFs or components. This decomposition is definitely based on the local spatial scales of the texture. However, RBF-linear, RBF-multiquadric,

24 332 S. M. A. Bhuiyan et al. (d) RBF-multiquadric (e) RBF-inverse multiquadric Fig. 14. (Continued ) and RBF-exponential interpolation methods result in more than three components as seen in Fig. 11 and Table 4, making them different from the actual synthetic texture components. From Fig. 11 it is further observed that the RBF-linear fails to produce meaningful components. These components are not enough orthogonal as well, which can be understood from the corresponding OIs in Table 4. Although the RBF-multiquadric interpolator produces some kind of meaningful components, its performance is not as good as RBF-cubic, RBF-thin-plate spline, RBF-inverse multiquadric, or RBF-Gaussian. Indeed, from the simulation with synthetic texture

25 Bidimensional Empirical Mode Decomposition 333 (f) RBF-exponential Fig. 14. (Continued ) image, RBF-cubic, RBF-thin-plate spline, RBF-inverse multiquadric, and RBF- Gaussian interpolators appear to be effective and/or the preferred methods for BEMD. Although the total number of iterations needed for the entire decomposition process is a factor indicating the computation involved, the number of iterations for the first few IMFs is more significant in this regard. Generally, the first iteration for the first IMF requires more computation and time than the first iteration for the second IMF, and so on. On the other hand, in general, the first IMF requires more iterations than the subsequent IMFs. This can be attributed to the fact that the number of extrema in the corresponding source images (alternately known in the

26 334 S. M. A. Bhuiyan et al. (g) RBF-Gaussian Fig. 14. (Continued ) Table 12. Comparison among different RBF interpolators with respect to D18 image of Fig. 13: total IMFs, iterations, and OI. RBF interpolators Total Iterations needed OI number of components Total IMF 1 IMF 2 IMF 3 IMF 4 IMF 5 Linear Cubic Thin-plate spline Multiquadric Inverse multiquadric Exponential Gaussian literature as intermediate residues) of subsequent modes decreases, which further affects the interpolation algorithm. Table 5 shows that IMF-1 needs the lowest number of iterations for RBF-Gaussian, while IMF-2 needs the lowest number of iterations for RBF-cubic; and IMF-1 needs the highest number of iterations for RBF-cubic, while IMF-2 needs the highest number of iterations for RBF-thin-plate spline. Achieved SD is an indication of local mean of the IMFs, where a lower SD indicates that the local mean is closer to zero. Since, for the synthetic texture, the desired SD threshold is achieved with all the RBF interpolation methods, as can be seen in Table 6, no conclusion can be drawn from the achieved SD based only on the synthetic texture decomposition. Although zero global mean does not necessarily indicate zero local mean, the global means of the decomposed components are given in Table 7 to compare them with the original components.

27 Bidimensional Empirical Mode Decomposition 335 Table 13. Comparison among different RBF interpolators with respect to D27 image of Fig. 13: total IMFs, iterations, and OI. RBF interpolators Total Iterations needed OI number of components Total IMF 1 IMF 2 IMF 3 IMF 4 IMF 5 Linear Cubic Thin-plate spline Multiquadric Inverse multiquadric Exponential Gaussian MSE and MAE between the decomposed components and the corresponding original components, as well as between their sum and the synthetic texture for the four preferred RBF interpolation methods are given in Tables 8 and 9. These tables show that for IMF-1 RBF-Gaussian causes the lowest MSE or MAE, while RBF-thin-plate spline causes the highest MSE or MAE. For IMF-2 and residue, RBF-cubic causes the lowest MSE or MAE. RBF-inverse multiquadric causes the highest MSE or MAE for IMF-2, while RBF-Gaussian causes the highest MSE or MAE for the residue. MSE or MAE is almost zero for the sum of decomposed components with all the RBF interpolators. Correlation coefficient (CC) may be a good measure to evaluate the match between the original components and the decomposed components. All the CC values as observed in Table 10 are good enough for the four preferred interpolation methods indicating successful decomposition into nearly original components. However, the performance of RBF-cubic may be better than the others. PSNR provides similar information about the effect of interpolation methods, as do the MSE and MAE. From Tables 4 to 11 and from Figs. 11 and 12, it appears that the RBF-cubic performs better than all other studied interpolators for this particular synthetic texture. Like the synthetic case, the total number of decomposed components, total iterations, and OIs are used for performance evaluation of BEMD with real texture images for different interpolation methods. As sample cases, the number of iterations for the first five IMFs are also observed for comparison based on the fact that the first few IMFs generally require more iterations and computation than the later IMFs. It has already been mentioned that the first 2D IMF provides the highest local spatial scales (finest local variation) of the image, while the final 2D IMF provides the lowest local spatial scales (coarsest local variation) of the image; and the final residue gives the trend of the image. Thus IMFs should always have some resemblance to the image at different scales of local spatial variation. Hence, visually evaluating the IMFs and matching them with the actual image provides additional performance measures. The simulation results in terms of the various performance indices given in Tables 12 and 13, and the characteristics of the IMFs obtained for the real texture images indicate that the effects of different interpolation methods

28 336 S. M. A. Bhuiyan et al. are quite similar for the synthetic texture. For example, RBF-linear and RBFexponential interpolators result in a large number of IMFs with a high value of OI (lower local orthogonality). Beside these, the IMFs do not resemble the original textures. Hence, RBF-linear and RBF-exponential interpolators do not appear to be suitable for the BEMD process. The simulation using real texture images indicates that RBF-cubic, RBF-thin-plate spline, RBF-multiquadric, RBF-inverse multiquadric, and RBF-Gaussian interpolators can be good candidates for BEMD. However, simulation results from Tables 12 and 13 as well as visual evaluation of the 2D IMFs emphasize that RBF-cubic, and RBF-thin-plate spline interpolators may be preferable for all general-purpose decompositions using BEMD. Based on the characteristics of obtained interpolated surface with DT based interpolation methods, it has already been claimed in Sec that DT based interpolation methods are not suitable for BEMD unless additional manipulation is done. Careful observation of the envelope shapes shown in Fig. 10 and the envelope shapes for other images and/or source images at different iterations (not shown in this paper), and corresponding 2D IMFs indicates that the mean envelope having locally convex or concave like surface bounded by zero crossings may provide more meaningful decomposition along with higher local orthogonality (lower OI) and zero local mean of the components. Hence, the 2D surface interpolation methods producing locally convex or concave like smoother surfaces may be preferable. For this reason RBF-cubic, RBF-thin-plate spline, RBF-inverse multiquadric, and RBF- Gaussian interpolators provide better results than the others. This fact may be attributed to the required IMF characteristic as defined for 1D IMFs. 1 However, further investigation may be required to find the true reason for the success of a particular interpolation method in BEMD. 6. Conclusion In this paper, comprehensive experiments on BEMD have been made using DT and RBF based interpolation methods. The purpose of the analyses is to find the effect of interpolation methods on the performance of BEMD and to determine some preferable methods for the BEMD process. In the analysis, the emphasis is mainly given on the interpolation methods to govern the decomposition by providing less control over the other parameters. It has been observed that RBF-cubic, RBF-thinplate spline, RBF-inverse multiquadric, and RBF-Gaussian interpolators provide better results than the others, which is clearly supported by the simulation using both synthetic and real images in this paper. However, it may be difficult to reach a conclusion about the superiority of one particular interpolation method for BEMD, based on the experiments on only a few images, but it certainly sheds light on a few methods performing better than the others. On the other hand, it is believed that the results in this paper may be helpful to select a particular interpolation method depending on the specific needs, for example, local orthogonality, local mean, and total number of IMFs.

29 Bidimensional Empirical Mode Decomposition 337 References 1. N. E. Huang et al., The empirical mode decomposition and the Hilbert spectrum for non-linear and non-stationary time series analysis, Proc. Roy. Soc. Lond. A 454 (1998) N. E. Huang, Z. Shen and S. R. Long, A new view of nonlinear water waves: the Hilbert spectrum, Ann. Rev. Fluid Mech. 31 (1999) J. C. Nunes, Y. Bouaoune, E. Delechelle, O. Niang and Ph. Bunel, Image analysis by bidimensional empirical mode decomposition, Image Vision Comput. 21 (2003) J. C. Nunes, S. Guyot and E. Deléchelle, Texture analysis based on local analysis of the bidimensional empirical mode decomposition, Mach.Vis.Appl.16(3) (2005) A. Linderhed, 2D empirical mode decompositions in the spirit of image compression, Proc. SPIE 4738 (2002) C. Damerval, S. Meignen and V. Perrier, A fast algorithm for bidimensional EMD, IEEE Signal Process. Lett. 12(10) (2005) Y. Xu, B. Liu and S. Riemenschneider, Two-dimensional empirical mode decomposition by finite elements, Proc. Roy. Soc. Lond. A 462 (2006) Z. Liu, H. Wang and S. Peng, Texture classification through empirical mode decomposition, Proc. 17th Int. Conf. Pattern Recognition ICRP 2004 (Cambridge UK, August 2004), pp Z. Liu, H. Wang and S. Peng, Texture segmentation using directional empirical mode decomposition, Proc Int. Conf. Image Processing (ICIP-2004) (Singapore, 2004), pp S. R. Long, Applications of HHT in image analysis, in Hilbert Huang Transform and Its Applications, eds. N. E. Huang and S. S. P. Shen (World Scientific, 2005). 11. Z. Liu and S. Peng, Boundary processing of bidimensional EMD using texture synthesis, IEEE Signal Process. Lett. 12(1) (2005) K. Zeng and M.-X. He, A simple boundary process technique for empirical mode decomposition, Proc IEEE Int. Geoscience and Remote Sensing Symposium (IGARSS 2004), Vol. 6 (2004), pp A. Linderhed, Variable sampling of the empirical mode decomposition of twodimensional signals, Int. J. Wavelet. Multiresol. Inform. Process. Special Issue on Frame Theory and Sampling Problems in Time-Frequency Analysis and Wavelet Theory (Part II) 3(3) (2005) A. Linderhed, Adaptive image compression with wavelet packets and empirical mode decomposition, Ph.D. Dissertation, Linköping Studies in Science and Technology (ISBN ). 15. B. Shen, Estimating the instantaneous frequencies of a multicomponent AM FM image by bidimensional empirical mode decomposition, Proc IEEE Int. Workshop on Intelligent Signal Processing (WISP 2005) (2005), pp N. E. Huang et al., A confidence limit for the empirical mode decomposition and Hilbert spectral analysis, Proc. Roy. Soc. Lond. A 459 (2003) G. Rilling, P. Flandrin and P. Goncalves, On empirical mode decomposition and its algorithms, Proc. IEEE-EURASIP Workshop on Nonlinear Signal and Image Processing, R. Franke and G. M. Nielson, Scattered data interpolation and applications: a tutorial and survey, in Geometric Modeling: Methods and Their Application, eds. H. Hagen and D. Roller (Springer, Berlin, 1991), pp

30 338 S. M. A. Bhuiyan et al. 19. G. M. Nielson, Scattered data modeling, IEEE Computer Graphics and Applications, 13(1) (1993) I. Amidror, Scattered data interpolation methods for electronic imaging systems: a survey, J. Elect. Imag. 11(2) (2002) R. L. Hardy, Multiquadratic equations of topography and other irregular surfaces, J. Geophys. Res. 76(8) (1971) M. D. Buhmann, Radial basis functions, Acta Numer. 9 (2000) J. C. Carr, W. R. Fright and R. K. Beatson, Surface interpolation with radial basis functions for medical imaging, IEEE Trans. Med. Imag. 16(1) (1997) J. Duchon, Splines minimizing rotation-invariant seminorms in Sobolev spaces, in Constructive Theory of Functions of Several Variables, eds. W. Schempp and K. Zeller (Springer, Berlin, 1977), pp F. L. Bookstein, Principal warps: thin-plate splines and the decomposition of deformations, IEEE Trans. Pattern Anal. Mach. Intell. 11(6) (1989) S. Rippa, An algorithm for selecting a good value for the parameter c in radial basis function interpolation, Adv. Comput. Math. 11 (1999) F. Aurenhammer and R. Klein, Voronoi diagrams, in Handbook of Computational Geometry, eds. J. R. Sack and J. Urrutia (North-Holland, Amsterdam, Netherlands, 2000), pp C. B. Barber, D. P. Dobkin and H. T. Huhdanpaa, The quickhull algorithm for convex hulls, ACM Trans. Math. Software 22 (1996) T. Y. Yang, Finite Element Structural Analysis (Prentice-Hall, Englewood Cliffs, NJ, 1986), pp C. Junsheng, Y. Dejie and Y. Yu, Research on the intrinsic mode function (IMF) criterion in EMD method, Mech. Syst. Signal Process. 20(4) (2006) P. Brodatz, Textures A Photographic Album for Artists and Designers (Dover, New York, 1966).

Research Article Fast and Adaptive Bidimensional Empirical Mode Decomposition Using Order-Statistics Filter Based Envelope Estimation

Research Article Fast and Adaptive Bidimensional Empirical Mode Decomposition Using Order-Statistics Filter Based Envelope Estimation Hindawi Publishing Corporation EURASIP Journal on Advances in Signal Processing Volume 8, Article ID 72856, 8 pages doi:.55/8/72856 Research Article Fast and Adaptive Bidimensional Empirical Mode Decomposition

More information

Adaptive Boundary Effect Processing For Empirical Mode Decomposition Using Template Matching

Adaptive Boundary Effect Processing For Empirical Mode Decomposition Using Template Matching Appl. Math. Inf. Sci. 7, No. 1L, 61-66 (2013) 61 Applied Mathematics & Information Sciences An International Journal Adaptive Boundary Effect Processing For Empirical Mode Decomposition Using Template

More information

An Improved Images Watermarking Scheme Using FABEMD Decomposition and DCT

An Improved Images Watermarking Scheme Using FABEMD Decomposition and DCT An Improved Images Watermarking Scheme Using FABEMD Decomposition and DCT Noura Aherrahrou and Hamid Tairi University Sidi Mohamed Ben Abdellah, Faculty of Sciences, Dhar El mahraz, LIIAN, Department of

More information

Interpolation artifacts and bidimensional ensemble empirical mode decomposition

Interpolation artifacts and bidimensional ensemble empirical mode decomposition Interpolation artifacts and bidimensional ensemble empirical mode decomposition Jiajun Han* University of Alberta, Edmonton, Alberta, Canada, hjiajun@ualberta.ca Mirko van der Baan University of Alberta,

More information

Empirical Mode Decomposition Based Denoising by Customized Thresholding

Empirical Mode Decomposition Based Denoising by Customized Thresholding Vol:11, No:5, 17 Empirical Mode Decomposition Based Denoising by Customized Thresholding Wahiba Mohguen, Raïs El hadi Bekka International Science Index, Electronics and Communication Engineering Vol:11,

More information

Equivalent Effect Function and Fast Intrinsic Mode Decomposition

Equivalent Effect Function and Fast Intrinsic Mode Decomposition Equivalent Effect Function and Fast Intrinsic Mode Decomposition Louis Yu Lu E-mail: louisyulu@gmail.com Abstract: The Equivalent Effect Function (EEF) is defined as having the identical integral values

More information

FPGA Implementation of HHT for Feature Extraction of Signals

FPGA Implementation of HHT for Feature Extraction of Signals IJSTE - International Journal of Science Technology & Engineering Volume 3 Issue 01 July 2016 ISSN (online): 2349-784X FPGA Implementation of HHT for Feature Extraction of Signals Mahesh Baban Shinde Manish

More information

Determining optimal value of the shape parameter c in RBF for unequal distances topographical points by Cross-Validation algorithm

Determining optimal value of the shape parameter c in RBF for unequal distances topographical points by Cross-Validation algorithm Journal of Mathematical Modeling Vol. 5, No. 1, 2017, pp. 53-60 JMM Determining optimal value of the shape parameter c in RBF for unequal distances topographical points by Cross-Validation algorithm Mohammadreza

More information

Separation of Surface Roughness Profile from Raw Contour based on Empirical Mode Decomposition Shoubin LIU 1, a*, Hui ZHANG 2, b

Separation of Surface Roughness Profile from Raw Contour based on Empirical Mode Decomposition Shoubin LIU 1, a*, Hui ZHANG 2, b International Conference on Advances in Mechanical Engineering and Industrial Informatics (AMEII 2015) Separation of Surface Roughness Profile from Raw Contour based on Empirical Mode Decomposition Shoubin

More information

Key words: B- Spline filters, filter banks, sub band coding, Pre processing, Image Averaging IJSER

Key words: B- Spline filters, filter banks, sub band coding, Pre processing, Image Averaging IJSER International Journal of Scientific & Engineering Research, Volume 7, Issue 9, September-2016 470 Analyzing Low Bit Rate Image Compression Using Filters and Pre Filtering PNV ABHISHEK 1, U VINOD KUMAR

More information

A Novel Triangle-based Method for Scattered Data Interpolation

A Novel Triangle-based Method for Scattered Data Interpolation Applied Mathematical Sciences Vol. 8, 24, no. 34, 677-6724 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/.2988/ams.24.49686 A Novel Triangle-based Method for Scattered Data Interpolation Salvatore Cuomo,

More information

Geostatistics 2D GMS 7.0 TUTORIALS. 1 Introduction. 1.1 Contents

Geostatistics 2D GMS 7.0 TUTORIALS. 1 Introduction. 1.1 Contents GMS 7.0 TUTORIALS 1 Introduction Two-dimensional geostatistics (interpolation) can be performed in GMS using the 2D Scatter Point module. The module is used to interpolate from sets of 2D scatter points

More information

Accurate Thermo-Fluid Simulation in Real Time Environments. Silvia Poles, Alberto Deponti, EnginSoft S.p.A. Frank Rhodes, Mentor Graphics

Accurate Thermo-Fluid Simulation in Real Time Environments. Silvia Poles, Alberto Deponti, EnginSoft S.p.A. Frank Rhodes, Mentor Graphics Accurate Thermo-Fluid Simulation in Real Time Environments Silvia Poles, Alberto Deponti, EnginSoft S.p.A. Frank Rhodes, Mentor Graphics M e c h a n i c a l a n a l y s i s W h i t e P a p e r w w w. m

More information

LATEST TRENDS on APPLIED MATHEMATICS, SIMULATION, MODELLING

LATEST TRENDS on APPLIED MATHEMATICS, SIMULATION, MODELLING 3D surface reconstruction of objects by using stereoscopic viewing Baki Koyuncu, Kurtuluş Küllü bkoyuncu@ankara.edu.tr kkullu@eng.ankara.edu.tr Computer Engineering Department, Ankara University, Ankara,

More information

pyeemd Documentation Release Perttu Luukko

pyeemd Documentation Release Perttu Luukko pyeemd Documentation Release 1.3.1 Perttu Luukko August 10, 2016 Contents 1 Contents: 3 1.1 Installing pyeemd............................................ 3 1.2 Tutorial..................................................

More information

USING LINEAR PREDICTION TO MITIGATE END EFFECTS IN EMPIRICAL MODE DECOMPOSITION. Steven Sandoval, Matthew Bredin, and Phillip L.

USING LINEAR PREDICTION TO MITIGATE END EFFECTS IN EMPIRICAL MODE DECOMPOSITION. Steven Sandoval, Matthew Bredin, and Phillip L. USING LINEAR PREDICTION TO MITIGATE END EFFECTS IN EMPIRICAL MODE DECOMPOSITION Steven Sandoval, Matthew Bredin, and Phillip L. De Leon New Mexico State University Klipsch School of Electrical and Computer

More information

DIGITAL TERRAIN MODELS

DIGITAL TERRAIN MODELS DIGITAL TERRAIN MODELS 1 Digital Terrain Models Dr. Mohsen Mostafa Hassan Badawy Remote Sensing Center GENERAL: A Digital Terrain Models (DTM) is defined as the digital representation of the spatial distribution

More information

Image Enhancement Techniques for Fingerprint Identification

Image Enhancement Techniques for Fingerprint Identification March 2013 1 Image Enhancement Techniques for Fingerprint Identification Pankaj Deshmukh, Siraj Pathan, Riyaz Pathan Abstract The aim of this paper is to propose a new method in fingerprint enhancement

More information

Learning based face hallucination techniques: A survey

Learning based face hallucination techniques: A survey Vol. 3 (2014-15) pp. 37-45. : A survey Premitha Premnath K Department of Computer Science & Engineering Vidya Academy of Science & Technology Thrissur - 680501, Kerala, India (email: premithakpnath@gmail.com)

More information

Parameterization of Triangular Meshes with Virtual Boundaries

Parameterization of Triangular Meshes with Virtual Boundaries Parameterization of Triangular Meshes with Virtual Boundaries Yunjin Lee 1;Λ Hyoung Seok Kim 2;y Seungyong Lee 1;z 1 Department of Computer Science and Engineering Pohang University of Science and Technology

More information

Image Transformation Techniques Dr. Rajeev Srivastava Dept. of Computer Engineering, ITBHU, Varanasi

Image Transformation Techniques Dr. Rajeev Srivastava Dept. of Computer Engineering, ITBHU, Varanasi Image Transformation Techniques Dr. Rajeev Srivastava Dept. of Computer Engineering, ITBHU, Varanasi 1. Introduction The choice of a particular transform in a given application depends on the amount of

More information

Edge and local feature detection - 2. Importance of edge detection in computer vision

Edge and local feature detection - 2. Importance of edge detection in computer vision Edge and local feature detection Gradient based edge detection Edge detection by function fitting Second derivative edge detectors Edge linking and the construction of the chain graph Edge and local feature

More information

SOME stereo image-matching methods require a user-selected

SOME stereo image-matching methods require a user-selected IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, VOL. 3, NO. 2, APRIL 2006 207 Seed Point Selection Method for Triangle Constrained Image Matching Propagation Qing Zhu, Bo Wu, and Zhi-Xiang Xu Abstract In order

More information

Comparing different interpolation methods on two-dimensional test functions

Comparing different interpolation methods on two-dimensional test functions Comparing different interpolation methods on two-dimensional test functions Thomas Mühlenstädt, Sonja Kuhnt May 28, 2009 Keywords: Interpolation, computer experiment, Kriging, Kernel interpolation, Thin

More information

Fuzzy C-means with Bi-dimensional Empirical Mode Decomposition for Segmentation of Microarray Image

Fuzzy C-means with Bi-dimensional Empirical Mode Decomposition for Segmentation of Microarray Image www.ijcsi.org 316 Fuzzy C-means with Bi-dimensional Empirical Mode Decomposition for Segmentation of Microarray Image J.Harikiran 1, D.RamaKrishna 2, M.L.Phanendra 3, Dr.P.V.Lakshmi 4, Dr.R.Kiran Kumar

More information

arxiv: v1 [cs.cv] 2 May 2016

arxiv: v1 [cs.cv] 2 May 2016 16-811 Math Fundamentals for Robotics Comparison of Optimization Methods in Optical Flow Estimation Final Report, Fall 2015 arxiv:1605.00572v1 [cs.cv] 2 May 2016 Contents Noranart Vesdapunt Master of Computer

More information

Adaptive osculatory rational interpolation for image processing

Adaptive osculatory rational interpolation for image processing Journal of Computational and Applied Mathematics 195 (2006) 46 53 www.elsevier.com/locate/cam Adaptive osculatory rational interpolation for image processing Min Hu a, Jieqing Tan b, a College of Computer

More information

A Modified Spline Interpolation Method for Function Reconstruction from Its Zero-Crossings

A Modified Spline Interpolation Method for Function Reconstruction from Its Zero-Crossings Scientific Papers, University of Latvia, 2010. Vol. 756 Computer Science and Information Technologies 207 220 P. A Modified Spline Interpolation Method for Function Reconstruction from Its Zero-Crossings

More information

Fast and Adaptive Bidimensional Empirical Mode Decomposition for the Real-time Video Fusion

Fast and Adaptive Bidimensional Empirical Mode Decomposition for the Real-time Video Fusion Fast and Adaptive Bidimensional Empirical Mode Decomposition for the Real-time Video Fusion Maciej Wielgus Institute of Micromechanics and Photonics Warsaw University of Technology Warsaw, Poland maciek.wielgus@gmail.com

More information

ISSN (ONLINE): , VOLUME-3, ISSUE-1,

ISSN (ONLINE): , VOLUME-3, ISSUE-1, PERFORMANCE ANALYSIS OF LOSSLESS COMPRESSION TECHNIQUES TO INVESTIGATE THE OPTIMUM IMAGE COMPRESSION TECHNIQUE Dr. S. Swapna Rani Associate Professor, ECE Department M.V.S.R Engineering College, Nadergul,

More information

New structural similarity measure for image comparison

New structural similarity measure for image comparison University of Wollongong Research Online Faculty of Engineering and Information Sciences - Papers: Part A Faculty of Engineering and Information Sciences 2012 New structural similarity measure for image

More information

WEINER FILTER AND SUB-BLOCK DECOMPOSITION BASED IMAGE RESTORATION FOR MEDICAL APPLICATIONS

WEINER FILTER AND SUB-BLOCK DECOMPOSITION BASED IMAGE RESTORATION FOR MEDICAL APPLICATIONS WEINER FILTER AND SUB-BLOCK DECOMPOSITION BASED IMAGE RESTORATION FOR MEDICAL APPLICATIONS ARIFA SULTANA 1 & KANDARPA KUMAR SARMA 2 1,2 Department of Electronics and Communication Engineering, Gauhati

More information

Coarse-to-fine image registration

Coarse-to-fine image registration Today we will look at a few important topics in scale space in computer vision, in particular, coarseto-fine approaches, and the SIFT feature descriptor. I will present only the main ideas here to give

More information

CHAPTER 3 WAVELET DECOMPOSITION USING HAAR WAVELET

CHAPTER 3 WAVELET DECOMPOSITION USING HAAR WAVELET 69 CHAPTER 3 WAVELET DECOMPOSITION USING HAAR WAVELET 3.1 WAVELET Wavelet as a subject is highly interdisciplinary and it draws in crucial ways on ideas from the outside world. The working of wavelet in

More information

EE795: Computer Vision and Intelligent Systems

EE795: Computer Vision and Intelligent Systems EE795: Computer Vision and Intelligent Systems Spring 2012 TTh 17:30-18:45 FDH 204 Lecture 14 130307 http://www.ee.unlv.edu/~b1morris/ecg795/ 2 Outline Review Stereo Dense Motion Estimation Translational

More information

Möbius Transformations in Scientific Computing. David Eppstein

Möbius Transformations in Scientific Computing. David Eppstein Möbius Transformations in Scientific Computing David Eppstein Univ. of California, Irvine School of Information and Computer Science (including joint work with Marshall Bern from WADS 01 and SODA 03) Outline

More information

Recent Developments in Model-based Derivative-free Optimization

Recent Developments in Model-based Derivative-free Optimization Recent Developments in Model-based Derivative-free Optimization Seppo Pulkkinen April 23, 2010 Introduction Problem definition The problem we are considering is a nonlinear optimization problem with constraints:

More information

Digital Image Processing

Digital Image Processing Digital Image Processing Third Edition Rafael C. Gonzalez University of Tennessee Richard E. Woods MedData Interactive PEARSON Prentice Hall Pearson Education International Contents Preface xv Acknowledgments

More information

MRI Brain Image Segmentation Using an AM-FM Model

MRI Brain Image Segmentation Using an AM-FM Model MRI Brain Image Segmentation Using an AM-FM Model Marios S. Pattichis', Helen Petropoulos2, and William M. Brooks2 1 Department of Electrical and Computer Engineering, The University of New Mexico, Albuquerque,

More information

Image Coding with Active Appearance Models

Image Coding with Active Appearance Models Image Coding with Active Appearance Models Simon Baker, Iain Matthews, and Jeff Schneider CMU-RI-TR-03-13 The Robotics Institute Carnegie Mellon University Abstract Image coding is the task of representing

More information

Image denoising in the wavelet domain using Improved Neigh-shrink

Image denoising in the wavelet domain using Improved Neigh-shrink Image denoising in the wavelet domain using Improved Neigh-shrink Rahim Kamran 1, Mehdi Nasri, Hossein Nezamabadi-pour 3, Saeid Saryazdi 4 1 Rahimkamran008@gmail.com nasri_me@yahoo.com 3 nezam@uk.ac.ir

More information

Module 1 Lecture Notes 2. Optimization Problem and Model Formulation

Module 1 Lecture Notes 2. Optimization Problem and Model Formulation Optimization Methods: Introduction and Basic concepts 1 Module 1 Lecture Notes 2 Optimization Problem and Model Formulation Introduction In the previous lecture we studied the evolution of optimization

More information

Evaluation of texture features for image segmentation

Evaluation of texture features for image segmentation RIT Scholar Works Articles 9-14-2001 Evaluation of texture features for image segmentation Navid Serrano Jiebo Luo Andreas Savakis Follow this and additional works at: http://scholarworks.rit.edu/article

More information

Two Dimensional Microwave Imaging Using a Divide and Unite Algorithm

Two Dimensional Microwave Imaging Using a Divide and Unite Algorithm Two Dimensional Microwave Imaging Using a Divide and Unite Algorithm Disha Shur 1, K. Yaswanth 2, and Uday K. Khankhoje 2 1 Indian Institute of Engineering Science and Technology, Shibpur, India 2 Indian

More information

EMPIRICAL MODE DECOMPOSITION: APPLICATIONS ON SIGNAL AND IMAGE PROCESSING

EMPIRICAL MODE DECOMPOSITION: APPLICATIONS ON SIGNAL AND IMAGE PROCESSING Advances in Adaptive Data Analysis Vol. 1, No. 1 (2009) 125 175 c World Scientific Publishing Company EMPIRICAL MODE DECOMPOSITION: APPLICATIONS ON SIGNAL AND IMAGE PROCESSING JEAN-CLAUDE NUNES and ÉRIC

More information

Preferred directions for resolving the non-uniqueness of Delaunay triangulations

Preferred directions for resolving the non-uniqueness of Delaunay triangulations Preferred directions for resolving the non-uniqueness of Delaunay triangulations Christopher Dyken and Michael S. Floater Abstract: This note proposes a simple rule to determine a unique triangulation

More information

CPSC 695. Methods for interpolation and analysis of continuing surfaces in GIS Dr. M. Gavrilova

CPSC 695. Methods for interpolation and analysis of continuing surfaces in GIS Dr. M. Gavrilova CPSC 695 Methods for interpolation and analysis of continuing surfaces in GIS Dr. M. Gavrilova Overview Data sampling for continuous surfaces Interpolation methods Global interpolation Local interpolation

More information

Flexible-Hybrid Sequential Floating Search in Statistical Feature Selection

Flexible-Hybrid Sequential Floating Search in Statistical Feature Selection Flexible-Hybrid Sequential Floating Search in Statistical Feature Selection Petr Somol 1,2, Jana Novovičová 1,2, and Pavel Pudil 2,1 1 Dept. of Pattern Recognition, Institute of Information Theory and

More information

Adaptive Wavelet Image Denoising Based on the Entropy of Homogenus Regions

Adaptive Wavelet Image Denoising Based on the Entropy of Homogenus Regions International Journal of Electrical and Electronic Science 206; 3(4): 9-25 http://www.aascit.org/journal/ijees ISSN: 2375-2998 Adaptive Wavelet Image Denoising Based on the Entropy of Homogenus Regions

More information

Fairing Scalar Fields by Variational Modeling of Contours

Fairing Scalar Fields by Variational Modeling of Contours Fairing Scalar Fields by Variational Modeling of Contours Martin Bertram University of Kaiserslautern, Germany Abstract Volume rendering and isosurface extraction from three-dimensional scalar fields are

More information

DIGITAL TERRAIN MODELLING. Endre Katona University of Szeged Department of Informatics

DIGITAL TERRAIN MODELLING. Endre Katona University of Szeged Department of Informatics DIGITAL TERRAIN MODELLING Endre Katona University of Szeged Department of Informatics katona@inf.u-szeged.hu The problem: data sources data structures algorithms DTM = Digital Terrain Model Terrain function:

More information

Segmentation and Grouping

Segmentation and Grouping Segmentation and Grouping How and what do we see? Fundamental Problems ' Focus of attention, or grouping ' What subsets of pixels do we consider as possible objects? ' All connected subsets? ' Representation

More information

Efficient Image Compression of Medical Images Using the Wavelet Transform and Fuzzy c-means Clustering on Regions of Interest.

Efficient Image Compression of Medical Images Using the Wavelet Transform and Fuzzy c-means Clustering on Regions of Interest. Efficient Image Compression of Medical Images Using the Wavelet Transform and Fuzzy c-means Clustering on Regions of Interest. D.A. Karras, S.A. Karkanis and D. E. Maroulis University of Piraeus, Dept.

More information

EECS 556 Image Processing W 09. Interpolation. Interpolation techniques B splines

EECS 556 Image Processing W 09. Interpolation. Interpolation techniques B splines EECS 556 Image Processing W 09 Interpolation Interpolation techniques B splines What is image processing? Image processing is the application of 2D signal processing methods to images Image representation

More information

Surfaces, meshes, and topology

Surfaces, meshes, and topology Surfaces from Point Samples Surfaces, meshes, and topology A surface is a 2-manifold embedded in 3- dimensional Euclidean space Such surfaces are often approximated by triangle meshes 2 1 Triangle mesh

More information

Facial Expression Morphing and Animation with Local Warping Methods

Facial Expression Morphing and Animation with Local Warping Methods Facial Expression Morphing and Animation with Local Warping Methods Daw-Tung Lin and Han Huang Department of Computer Science and Information Engineering Chung Hua University 30 Tung-shiang, Hsin-chu,

More information

A new 8-node quadrilateral spline finite element

A new 8-node quadrilateral spline finite element Journal of Computational and Applied Mathematics 195 (2006) 54 65 www.elsevier.com/locate/cam A new 8-node quadrilateral spline finite element Chong-Jun Li, Ren-Hong Wang Institute of Mathematical Sciences,

More information

Periodicity Extraction using Superposition of Distance Matching Function and One-dimensional Haar Wavelet Transform

Periodicity Extraction using Superposition of Distance Matching Function and One-dimensional Haar Wavelet Transform Periodicity Extraction using Superposition of Distance Matching Function and One-dimensional Haar Wavelet Transform Dr. N.U. Bhajantri Department of Computer Science & Engineering, Government Engineering

More information

PROPERTIES OF NATURAL ELEMENT COORDINATES ON ANY POLYHEDRON

PROPERTIES OF NATURAL ELEMENT COORDINATES ON ANY POLYHEDRON PROPRTIS OF NATURAL LMNT COORDINATS ON ANY POLYHDRON P. Milbradt and T. Fröbel Institute of Computer Science in Civil ngineering, Univercity of Hanover, 3067, Hanover, Germany; PH (+49) 5-76-5757; FAX

More information

Image Inpainting Using Sparsity of the Transform Domain

Image Inpainting Using Sparsity of the Transform Domain Image Inpainting Using Sparsity of the Transform Domain H. Hosseini*, N.B. Marvasti, Student Member, IEEE, F. Marvasti, Senior Member, IEEE Advanced Communication Research Institute (ACRI) Department of

More information

Approximation of a Fuzzy Function by Using Radial Basis Functions Interpolation

Approximation of a Fuzzy Function by Using Radial Basis Functions Interpolation International Journal of Mathematical Modelling & Computations Vol. 07, No. 03, Summer 2017, 299-307 Approximation of a Fuzzy Function by Using Radial Basis Functions Interpolation R. Firouzdor a and M.

More information

Cluster EMD and its Statistical Application

Cluster EMD and its Statistical Application Cluster EMD and its Statistical Application Donghoh Kim and Heeseok Oh Sejong University and Seoul National University November 10, 2007 1/27 Contents 1. Multi-scale Concept 2. Decomposition 3. Cluster

More information

Compression of RADARSAT Data with Block Adaptive Wavelets Abstract: 1. Introduction

Compression of RADARSAT Data with Block Adaptive Wavelets Abstract: 1. Introduction Compression of RADARSAT Data with Block Adaptive Wavelets Ian Cumming and Jing Wang Department of Electrical and Computer Engineering The University of British Columbia 2356 Main Mall, Vancouver, BC, Canada

More information

3D Mesh Sequence Compression Using Thin-plate Spline based Prediction

3D Mesh Sequence Compression Using Thin-plate Spline based Prediction Appl. Math. Inf. Sci. 10, No. 4, 1603-1608 (2016) 1603 Applied Mathematics & Information Sciences An International Journal http://dx.doi.org/10.18576/amis/100440 3D Mesh Sequence Compression Using Thin-plate

More information

Iterated Functions Systems and Fractal Coding

Iterated Functions Systems and Fractal Coding Qing Jun He 90121047 Math 308 Essay Iterated Functions Systems and Fractal Coding 1. Introduction Fractal coding techniques are based on the theory of Iterated Function Systems (IFS) founded by Hutchinson

More information

3 Nonlinear Regression

3 Nonlinear Regression 3 Linear models are often insufficient to capture the real-world phenomena. That is, the relation between the inputs and the outputs we want to be able to predict are not linear. As a consequence, nonlinear

More information

Image Denoising using SWT 2D Wavelet Transform

Image Denoising using SWT 2D Wavelet Transform IJSTE - International Journal of Science Technology & Engineering Volume 3 Issue 01 July 2016 ISSN (online): 2349-784X Image Denoising using SWT 2D Wavelet Transform Deep Singh Bedi Department of Electronics

More information

Algorithm That Mimics Human Perceptual Grouping of Dot Patterns

Algorithm That Mimics Human Perceptual Grouping of Dot Patterns Algorithm That Mimics Human Perceptual Grouping of Dot Patterns G. Papari and N. Petkov Institute of Mathematics and Computing Science, University of Groningen, P.O.Box 800, 9700 AV Groningen, The Netherlands

More information

WAVELET TRANSFORM BASED FEATURE DETECTION

WAVELET TRANSFORM BASED FEATURE DETECTION WAVELET TRANSFORM BASED FEATURE DETECTION David Bařina Doctoral Degree Programme (1), DCGM, FIT BUT E-mail: ibarina@fit.vutbr.cz Supervised by: Pavel Zemčík E-mail: zemcik@fit.vutbr.cz ABSTRACT This paper

More information

CHAPTER 2 TEXTURE CLASSIFICATION METHODS GRAY LEVEL CO-OCCURRENCE MATRIX AND TEXTURE UNIT

CHAPTER 2 TEXTURE CLASSIFICATION METHODS GRAY LEVEL CO-OCCURRENCE MATRIX AND TEXTURE UNIT CHAPTER 2 TEXTURE CLASSIFICATION METHODS GRAY LEVEL CO-OCCURRENCE MATRIX AND TEXTURE UNIT 2.1 BRIEF OUTLINE The classification of digital imagery is to extract useful thematic information which is one

More information

Estimating normal vectors and curvatures by centroid weights

Estimating normal vectors and curvatures by centroid weights Computer Aided Geometric Design 21 (2004) 447 458 www.elsevier.com/locate/cagd Estimating normal vectors and curvatures by centroid weights Sheng-Gwo Chen, Jyh-Yang Wu Department of Mathematics, National

More information

Texture Image Segmentation using FCM

Texture Image Segmentation using FCM Proceedings of 2012 4th International Conference on Machine Learning and Computing IPCSIT vol. 25 (2012) (2012) IACSIT Press, Singapore Texture Image Segmentation using FCM Kanchan S. Deshmukh + M.G.M

More information

Model-based segmentation and recognition from range data

Model-based segmentation and recognition from range data Model-based segmentation and recognition from range data Jan Boehm Institute for Photogrammetry Universität Stuttgart Germany Keywords: range image, segmentation, object recognition, CAD ABSTRACT This

More information

Empirical Mode Decomposition Analysis using Rational Splines

Empirical Mode Decomposition Analysis using Rational Splines Empirical Mode Decomposition Analysis using Rational Splines Geoff Pegram Pegram, GGS, MC Peel & TA McMahon, (28). Empirical Mode Decomposition using rational splines: an application to rainfall time series.

More information

ADVANCED IMAGE PROCESSING METHODS FOR ULTRASONIC NDE RESEARCH C. H. Chen, University of Massachusetts Dartmouth, N.

ADVANCED IMAGE PROCESSING METHODS FOR ULTRASONIC NDE RESEARCH C. H. Chen, University of Massachusetts Dartmouth, N. ADVANCED IMAGE PROCESSING METHODS FOR ULTRASONIC NDE RESEARCH C. H. Chen, University of Massachusetts Dartmouth, N. Dartmouth, MA USA Abstract: The significant progress in ultrasonic NDE systems has now

More information

Contents. I The Basic Framework for Stationary Problems 1

Contents. I The Basic Framework for Stationary Problems 1 page v Preface xiii I The Basic Framework for Stationary Problems 1 1 Some model PDEs 3 1.1 Laplace s equation; elliptic BVPs... 3 1.1.1 Physical experiments modeled by Laplace s equation... 5 1.2 Other

More information

3 Nonlinear Regression

3 Nonlinear Regression CSC 4 / CSC D / CSC C 3 Sometimes linear models are not sufficient to capture the real-world phenomena, and thus nonlinear models are necessary. In regression, all such models will have the same basic

More information

Local Image Registration: An Adaptive Filtering Framework

Local Image Registration: An Adaptive Filtering Framework Local Image Registration: An Adaptive Filtering Framework Gulcin Caner a,a.murattekalp a,b, Gaurav Sharma a and Wendi Heinzelman a a Electrical and Computer Engineering Dept.,University of Rochester, Rochester,

More information

PRINCIPAL COMPONENT ANALYSIS IMAGE DENOISING USING LOCAL PIXEL GROUPING

PRINCIPAL COMPONENT ANALYSIS IMAGE DENOISING USING LOCAL PIXEL GROUPING PRINCIPAL COMPONENT ANALYSIS IMAGE DENOISING USING LOCAL PIXEL GROUPING Divesh Kumar 1 and Dheeraj Kalra 2 1 Department of Electronics & Communication Engineering, IET, GLA University, Mathura 2 Department

More information

Audio Watermarking using Colour Image Based on EMD and DCT

Audio Watermarking using Colour Image Based on EMD and DCT Audio Watermarking using Colour Image Based on EMD and Suhail Yoosuf 1, Ann Mary Alex 2 P. G. Scholar, Department of Electronics and Communication, Mar Baselios College of Engineering and Technology, Trivandrum,

More information

Robust biometric image watermarking for fingerprint and face template protection

Robust biometric image watermarking for fingerprint and face template protection Robust biometric image watermarking for fingerprint and face template protection Mayank Vatsa 1, Richa Singh 1, Afzel Noore 1a),MaxM.Houck 2, and Keith Morris 2 1 West Virginia University, Morgantown,

More information

Occluded Facial Expression Tracking

Occluded Facial Expression Tracking Occluded Facial Expression Tracking Hugo Mercier 1, Julien Peyras 2, and Patrice Dalle 1 1 Institut de Recherche en Informatique de Toulouse 118, route de Narbonne, F-31062 Toulouse Cedex 9 2 Dipartimento

More information

Locally Weighted Least Squares Regression for Image Denoising, Reconstruction and Up-sampling

Locally Weighted Least Squares Regression for Image Denoising, Reconstruction and Up-sampling Locally Weighted Least Squares Regression for Image Denoising, Reconstruction and Up-sampling Moritz Baecher May 15, 29 1 Introduction Edge-preserving smoothing and super-resolution are classic and important

More information

The Anatomical Equivalence Class Formulation and its Application to Shape-based Computational Neuroanatomy

The Anatomical Equivalence Class Formulation and its Application to Shape-based Computational Neuroanatomy The Anatomical Equivalence Class Formulation and its Application to Shape-based Computational Neuroanatomy Sokratis K. Makrogiannis, PhD From post-doctoral research at SBIA lab, Department of Radiology,

More information

Interpreter aided salt boundary segmentation using SVM regression shape deformation technique

Interpreter aided salt boundary segmentation using SVM regression shape deformation technique Interpreter aided salt boundary segmentation using SVM regression shape deformation technique Problem Introduction For the seismic imaging project in offshore oil and gas exploration, a good velocity model

More information

Short Survey on Static Hand Gesture Recognition

Short Survey on Static Hand Gesture Recognition Short Survey on Static Hand Gesture Recognition Huu-Hung Huynh University of Science and Technology The University of Danang, Vietnam Duc-Hoang Vo University of Science and Technology The University of

More information

Line Segment Based Watershed Segmentation

Line Segment Based Watershed Segmentation Line Segment Based Watershed Segmentation Johan De Bock 1 and Wilfried Philips Dep. TELIN/TW07, Ghent University Sint-Pietersnieuwstraat 41, B-9000 Ghent, Belgium jdebock@telin.ugent.be Abstract. In this

More information

CHAPTER 6 MODIFIED FUZZY TECHNIQUES BASED IMAGE SEGMENTATION

CHAPTER 6 MODIFIED FUZZY TECHNIQUES BASED IMAGE SEGMENTATION CHAPTER 6 MODIFIED FUZZY TECHNIQUES BASED IMAGE SEGMENTATION 6.1 INTRODUCTION Fuzzy logic based computational techniques are becoming increasingly important in the medical image analysis arena. The significant

More information

IMPROVED MOTION-BASED LOCALIZED SUPER RESOLUTION TECHNIQUE USING DISCRETE WAVELET TRANSFORM FOR LOW RESOLUTION VIDEO ENHANCEMENT

IMPROVED MOTION-BASED LOCALIZED SUPER RESOLUTION TECHNIQUE USING DISCRETE WAVELET TRANSFORM FOR LOW RESOLUTION VIDEO ENHANCEMENT 17th European Signal Processing Conference (EUSIPCO 009) Glasgow, Scotland, August 4-8, 009 IMPROVED MOTION-BASED LOCALIZED SUPER RESOLUTION TECHNIQUE USING DISCRETE WAVELET TRANSFORM FOR LOW RESOLUTION

More information

EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm

EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm Group 1: Mina A. Makar Stanford University mamakar@stanford.edu Abstract In this report, we investigate the application of the Scale-Invariant

More information

Towards Automatic Recognition of Fonts using Genetic Approach

Towards Automatic Recognition of Fonts using Genetic Approach Towards Automatic Recognition of Fonts using Genetic Approach M. SARFRAZ Department of Information and Computer Science King Fahd University of Petroleum and Minerals KFUPM # 1510, Dhahran 31261, Saudi

More information

A MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING

A MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING Proceedings of the 1994 IEEE International Conference on Image Processing (ICIP-94), pp. 530-534. (Austin, Texas, 13-16 November 1994.) A MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING

More information

Denoising and Edge Detection Using Sobelmethod

Denoising and Edge Detection Using Sobelmethod International OPEN ACCESS Journal Of Modern Engineering Research (IJMER) Denoising and Edge Detection Using Sobelmethod P. Sravya 1, T. Rupa devi 2, M. Janardhana Rao 3, K. Sai Jagadeesh 4, K. Prasanna

More information

An Automated Image-based Method for Multi-Leaf Collimator Positioning Verification in Intensity Modulated Radiation Therapy

An Automated Image-based Method for Multi-Leaf Collimator Positioning Verification in Intensity Modulated Radiation Therapy An Automated Image-based Method for Multi-Leaf Collimator Positioning Verification in Intensity Modulated Radiation Therapy Chenyang Xu 1, Siemens Corporate Research, Inc., Princeton, NJ, USA Xiaolei Huang,

More information

03 - Reconstruction. Acknowledgements: Olga Sorkine-Hornung. CSCI-GA Geometric Modeling - Spring 17 - Daniele Panozzo

03 - Reconstruction. Acknowledgements: Olga Sorkine-Hornung. CSCI-GA Geometric Modeling - Spring 17 - Daniele Panozzo 3 - Reconstruction Acknowledgements: Olga Sorkine-Hornung Geometry Acquisition Pipeline Scanning: results in range images Registration: bring all range images to one coordinate system Stitching/ reconstruction:

More information

Face Tracking. Synonyms. Definition. Main Body Text. Amit K. Roy-Chowdhury and Yilei Xu. Facial Motion Estimation

Face Tracking. Synonyms. Definition. Main Body Text. Amit K. Roy-Chowdhury and Yilei Xu. Facial Motion Estimation Face Tracking Amit K. Roy-Chowdhury and Yilei Xu Department of Electrical Engineering, University of California, Riverside, CA 92521, USA {amitrc,yxu}@ee.ucr.edu Synonyms Facial Motion Estimation Definition

More information

Simple Mesh Examples to Illustrate Specific Finite Element Mesh Requirements

Simple Mesh Examples to Illustrate Specific Finite Element Mesh Requirements Simple Mesh Examples to Illustrate Specific Finite Element Mesh Requirements Peter Fleischmann, Robert Kosik, Bernhard Haindl, and Siegfried Selberherr Institute for Microelectronics, TU Vienna, Gu6hausstrafie

More information

PERFORMANCE ANALYSIS OF CANNY AND OTHER COMMONLY USED EDGE DETECTORS Sandeep Dhawan Director of Technology, OTTE, NEW YORK

PERFORMANCE ANALYSIS OF CANNY AND OTHER COMMONLY USED EDGE DETECTORS Sandeep Dhawan Director of Technology, OTTE, NEW YORK International Journal of Science, Environment and Technology, Vol. 3, No 5, 2014, 1759 1766 ISSN 2278-3687 (O) PERFORMANCE ANALYSIS OF CANNY AND OTHER COMMONLY USED EDGE DETECTORS Sandeep Dhawan Director

More information

13.472J/1.128J/2.158J/16.940J COMPUTATIONAL GEOMETRY

13.472J/1.128J/2.158J/16.940J COMPUTATIONAL GEOMETRY 13.472J/1.128J/2.158J/16.940J COMPUTATIONAL GEOMETRY Lecture 23 Dr. W. Cho Prof. N. M. Patrikalakis Copyright c 2003 Massachusetts Institute of Technology Contents 23 F.E. and B.E. Meshing Algorithms 2

More information

CHAPTER 8 COMPOUND CHARACTER RECOGNITION USING VARIOUS MODELS

CHAPTER 8 COMPOUND CHARACTER RECOGNITION USING VARIOUS MODELS CHAPTER 8 COMPOUND CHARACTER RECOGNITION USING VARIOUS MODELS 8.1 Introduction The recognition systems developed so far were for simple characters comprising of consonants and vowels. But there is one

More information