Image Mining by Matching Exemplars Using Entropy

Size: px
Start display at page:

Download "Image Mining by Matching Exemplars Using Entropy"

Transcription

1 Image Mining by Matching Exemplars Using Entropy Clark F. Olson University of Washington, Bothell Computing and Software Systems Campus Way NE, Campus Box Bothell, WA ABSTRACT This paper describes a methodology for off-line extraction of content from image and video libraries in order to support complex queries at run-time. The basic technique is to detect matches for exemplars in the images using a technique based on entropy that is invariant to many (and insensitive to others) of the imaging phenomena that can cause appearance-based techniques to fail. The technique has been successful in locating matches for exemplars even when the images are captured with completely different sensors (e.g. CCD versus FLIR). Current matching techniques are efficient for single images, but are likely to be prohibitively expensive for massive databases. We outline a hierarchical processing strategy that we believe has potential for searching such databases. Categories and Subject Descriptors H.3.7 [Information Systems]: Information Storage and Retrieval Digital Libraries General Terms Algorithms Keywords Data mining, image database, template matching, entropy 1. INTRODUCTION Many current image and video databases are databases in name only. They are unmanaged collections of data, rather than being managed by a database management system (DBMS) that provides a facility for forming and processing complex queries on the data. Considerable work on contentbased image retrieval has studied methods to extract general information from images in order to support some queries at run-time [20]. However, the generality of these methods precludes them from supporting specific, complex queries, such as In which video sequences does Tiger Woods appear after to Annika Sorenstam? or Which patients have MRI The copyright of these papers belongs to the paper s authors. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. MDM/KDD 03, August 27, 2003, Washington, DC, USA. images that contain two tumors of different types? The primary objective of this work is to develop algorithms and tools in order to create true databases from collections of multimedia data (specifically images and video sequences) by mining content from the data off-line in order to efficiently support complex queries at run-time. Shape-based image retrieval is a difficult problem. The reason for this is the intractability of the segmentation problem to extract shape information from complex image data. Extending such retrieval to video gains the benefit (and the curse) of orders of magnitude more data to examine. For this reason, we believe it is likely that the most common method for extracting content from image and video databases will be through matching against appearance-based exemplars. Appearance-based matching has achieved good results in many areas of computer vision, including face recognition [19] and object classification [11]. However, appearancebased techniques face difficulties when objects vary from the expected appearance due to differences in sensor, lighting, configuration, etc. We build upon previous work for robust image matching [13, 14] to extract matches in the database using image exemplars. Figure 1 shows an example of appearance-based recognition using an exemplar. In this example, a location in an orbital image is detected using an aerial image of the same location even though the images were captured with very different sensors, at different resolutions, and slightly different viewing direction. This example illustrates that successful extraction can be performed even when the exemplar has a different appearance from the image to which it is matched. The goal of this work is to develop a matching measure that is robust to changing sensors, lighting conditions, and image motion together with an efficient search strategy for locating good matches according to the measure. For example, this technique could be used to detect all instances of a particular location in a database of geographical images using an exemplar. Maximization of mutual information [10, 21] has been successful in robust matching. However, there is no efficient search strategy for large search spaces. In the above example (in which mutual information fails to find the correct result), matching was performed by transforming both the exemplar and the search image into new images by replacing each pixel with a local entropy measurement. The scoring function for each position of the exemplar examines how well

2 (a) (b) (c) Figure 1: Motivating example. (a) Orbital image of the Avawatz Mountains and Silurian Valley in California. (b) Aerial image showing a detail of (a). (c) Annotated image. The box shows the location of (b) in the orbital image. the local entropies are correlated between the exemplar and the corresponding locations in the search image. The optimal match was detected over affine transformations of the exemplar image. 2. RELATED WORK A large body of previous work exists on retrieving images satisfying various characteristics from an image collection [20]. Influential early systems include QBIC [5], Photobook [15], and the Virage engine [1]. The QBIC (Query By Image Content) system uses image frames and objects that have been segmented (manually or automatically) from an image. Images and objects segmented from them are represented by sets of features representing color, texture, shape, location, and sketch. Queries can be performed on entire images or individual objects using these features. Photobook similarly allows searches to be performed using extracted image features. Three types of image properties are used: appearance (for example, segmented faces), two-dimensional shape (tools, fish, etc.), and texture (wood, honeycomb, grass). Each of the image properties is represented in the database using a small set of coefficients (30 to 100) from which the original image can be reconstructed with little error. The image properties can also be combined to form searches that are more complex. Virage provides an image search engine has three functional parts: image analysis, image comparison, and management. Preprocessing operations and primitive extraction is performed during image analysis. During image comparison, similarity measures are applied to pairs of primitive vectors that have been extracted from images and combined using weights. The management function performs initialization, and manages the weights and primitive vector data. Recent systems include NeTra [9], which first performs segmentation of the image into distinct regions. Queries are performed through indexing using color, texture, shape, and spatial location. PicHunter [4] uses relevance feedback to predict the image desired by the user given their actions. The overall approach is Bayesian, with an explicit model of the user s actions. PicHunter uses annotated images; however, the annotations are hidden from the user to reduce confusion on the user s part. PicToSeek [6] uses color and shape invariants to retrieve images from a database. New color models are used to gain invariance to viewpoint, geometry, and illumination. Shape is extracted through color invariant edge detection. The shape and color invariants are combined into a single high-dimensional feature set for discriminating between objects. Note that such content-based image retrieval systems invariably require a user in the loop in order to refine queries and locate the desired images in the database. The techniques described here extract content from multimedia databases automatically. User involvement is required only for queries on the database, after the relevant information has been extracted from the multimedia data. 3. MATCHING WITH ENTROPY The basic technique that we use for detecting matches is to compare the local entropy against that of an exemplar image over some space of relative positions between the images [13]. For a discrete random variable A, with marginal probability distribution p A(a), the entropy is defined as: H(A) = p A(a)logp A(a). (1) a Note that 0 log 0 is taken to be zero, since: lim x log x =0. (2) x 0 We apply an entropy transformation to both the template and the reference image as follows. For each image location (x, y), we examine the intensity values in an image window (with some specified size k k) centered(x, y). The intensities are histogrammed and the entropy for the window is computed according to Eq. (1) above. In practice, it is useful to smooth the histograms prior to using Eq. (1). For efficiency, we use a histogram with 64 bins (6 bits of informa-

3 (a) (b) (c) (d) (e) Figure 2: Entropy images at different scales. (a) Original image. (b) Entropy image for 3 3 window. (c) Entropy image for 5 5 window. (d) Entropy image for window. (e) Entropy image for window. tion) and smooth the histogram using a Gaussian (σ =1.0 bins). Figure 2 shows an example of the entropy images generated at a variety of scales. It can be observed that the entropy transformation captures the amount of local variation at each location in the image. As the size of the window increases, the local entropies are spread and smoothed over a larger area. This property is useful in the generation of a multi-resolution search strategy. While these images are not invariant to all illumination changes and sensor characteristics, they are insensitive to many issues in multi-sensor matching while retaining much image information. The entropy images are invariant to bias in the original images and the correlation peaks are invariant to gain in the entropy images. Of course, illumination changes and different sensors cause more complex transformations than these, but we have found that the locations of the correlation peaks change little. Now, we want to maximize the normalized correlation between the entropy images over the search space. In general, this involves rotating and rescaling the template entropy image according to the pose parameters. (We can also include shearing and other deformations in the search space. Our examples use the six degree-of-freedom affine transformation.) The shape and scale of the template will, thus, vary over the search space and this must be accounted for in the normalization. In order to deal with this in the search for the best position, whenever we generate a rescaled template, we generate another template of the same shape storing the weight of each position in the correlation. Most positions will have a weight of one, but the edges of the template will have lower values, since bilinear interpolation does not have four pixels to interpolate from in this case. This template can be used in the Fast Fourier Transform (FFT) to quickly determine the normalization values for each translation of the template in the search strategy described below. 4. TOWARDS EFFICIENT SEARCH Since multimedia databases can store terabytes of information it is necessary to use highly efficient strategies for processing the data. In some cases it will be unacceptable for the mining to require days to operate, but it may be acceptable in large databases to mine the data over the course of hours (or overnight). Current strategies for robust appearance-based matching are computationally intensive. Matching using mutual information [10, 21] is a leading technique for such matching. However, developing a search strategy that is both general and efficient is difficult for this technique. Our current approach combines fast two-dimensional matching using the FFT, coarse sampling of the remaining pose parameters, and iterative refinement to detect the optimal position(s) [13]. Search over the two degrees-of-freedom corresponding to translation in the image can be performed very quickly in the frequency domain using the FFT, since this allows cross-correlation to be performed in O(n log n) time, where n is the size on the image. In more complex search spaces, including similarity and affine transformations, we sample the remaining parameters on a grid to ensure that a sample point is close to every point in the search space. For each sample point, the translation parameters are searched with the FFT and strong candidates are further refined using Powell s iterative optimization method [16]. This methodology yields good results for matching in a single image, but it is likely to be prohibitively expensive when matching against a very large database of images and video. A promising technique is guaranteed search for image matching [14, 18]. This method is able to prune large sections of the search space using little computation. These techniques have not, yet, been applied to image matching using statistical measures, such as mutual information or entropy matching. However, previous work on entropy matching [13] can be framed as a multi-resolution search problem, making such techniques feasible. Video databases with both spatial and temporal properties are even more complicated. One approach that has been suggested [3] is the use of 3D strings in the x, y, and time dimensions. Liu and Chen [8] use three-dimensional pattern matching techniques for this purpose. Their approach scales linearly with the number of video objects. In order to achieve efficient operations on very large databases, it will be necessary to obtain a sub-linear relationship with the number of video objects. This may be achieved using a hierarchy of video objects. The top of the hierarchy would be coarse, but would allow many video objects to be removed from consideration. Lower levels of the hierarchy would add detail and remove more video objects, until the remaining video objects could be examined explicitly. This methodology is similar to work in object recognition that uses a hierarchy of positions in the space of object po-

4 (a) (b) Figure 3: Location detection between aerial and orbital images. (a) Aerial image of California desert. (b) Detected location in an orbital image. sitions (i.e., the pose space) [2, 7, 12]. At the upper levels of the hierarchy, the method tries to prove that each cell of the pose space cannot contain a position meeting the search criteria. If successful for some cell, the entire cell can be removed from the search. Otherwise, the cell is subdivided into smaller cells (lower in the hierarchy) and the process is applied recursively. This technique succeeds since it never prunes a cell that could contain a position meeting the search criteria. Application of similar ideas to video databases has the potential to yield gains in efficiency when dealing with large databases. 5. RESULTS Figures 3-5 show examples of the application of our method to real images. In Fig. 3, the template is an image captured using a helicopter at an elevation of approximately 800 meters and the reference image is an orbital image that encompasses the same location. The correct location of the template is not easy to detect manually. Successful matching was performed with the affine transformation and the correct location was determined in 11.4 seconds on a 500 MHz workstation with a pixel template and a pixel reference image. Figure 4 shows an example where a military vehicle was detected in a CCD using an exemplar from an infrared image. We have applied these techniques also to images from Mars (Fig. 5). Both the template and the reference image were captured with the Imager for Mars Pathfinder (IMP) cameras. The images were captured at different times during the day, so the illumination and the shadowing of the terrain is different between the images. Furthermore, the reference image has undergone a non-linear warping operation to remove lens distortion, while the template has not. The matching algorithm was able to succeed despite these differences and required seconds with a pixel template and a pixel reference image for the problem in Fig FUTURE WORK In addition to research to find highly efficient search strategies for matching against exemplars, we have identified two other important areas for future work: highly robust tracking in video images and grouping of extracted objects. 6.1 Tracking Tracking objects in video is easier than detecting/recognizing them. However, highly robust methods are necessary to maintain query integrity. It is unacceptable for the tracking procedure to be fooled into locking onto a background object or drift off the correct target. Since the tracked object will change appearance over the course of the video sequence, thetemplateusedtotracktheobjectmustbeperiodically (or continually) updated. However, each time the template is updated, the target on which the template is centered drifts slightly from the desired position. For the tracking to be successful over a long sequence, localization must be performed with high precision each time the template is updated. One approach [17] is to compute the motion not just relative to the previous frame, but to use a set of previous frames. We intend to strengthen this approach using a robust template matching framework and high-precision subpixel estimation techniques [14]. Another important modification is to update the target template only when necessary, in order to reduce the opportunities for drift. The decision of when to update the target template can be determined using a quality of match measure between the current template and video frame. For example, we may decide that the template must be updated when

5 (a) (b) Figure 4: Recognition using entropy with an exemplar from a different sensor. (a) Tank exemplar from FLIR image. (b) Extracted tank location from a CCD image. (a) (b) Figure 5: Recognition using images taken with different filters and with different illumination. (a) CCD image of a Martian rock. (b) Detected location of rock in a CCD image with a different lens filter. an estimate of the standard deviation in the localization or the probability of failure has risen above a pre-determined value. 6.2 Grouping If objects are mined from video using exemplars, then the identity of the objects is known and they are trivially grouped according to the exemplar that generated them. On the other hand, techniques can be integrated into this framework that do not use exemplars. In addition, techniques for applications such as face recognition may use general face exemplars, without distinguishing between faces belonging to different people. For cases where there is no obvious partitioning of the detected objects, we wish to classify the objects according to some meaningful characteristic; in a video containing people, we should be able to detected cases where a particular face leaves the image sequence and subsequently returns. This requires grouping the detected face instances according to the appearance of each face, rather than classifying them as different objects. Our initial approach to matching separate face images, while allowing differences in lighting and affine transformation, uses techniques similar to the problem of extraction using exemplars [13, 14]. With this methodology, the extracted objects (such as faces) are treated as exemplars for matching against other extracted objects. For highly non-planar objects, this approach can be extended to use shape information inferred from the object class and image. For example, the three-dimensional structure of a face can be well approximated using a small number of parameters extracted from an image of the face. This shape information can used to improve the robustness of matching between two images of a face observed from different viewpoints, since a non-linear image warping making use of the shape information brings the images into better alignment. 7. SUMMARY Our goal in this paper is to demonstrate the data can be mined from image and video databases by performing matching against exemplars. The technique that we use for matching is based on comparing the local entropies in the images

6 at some relative position that can encompass translation, rotation, scaling, and skew. This allows matches to be detected even when the images are taken by different sensors and from different viewpoints. Search strategies for performing matching are currently efficient only for small sets of images. We have suggested techniques for future work that may lead to efficient search in large databases. 8. REFERENCES [1] J. R. Bach, C. Fuller, A. Gupta, A. Hampapur, B. Horowitz, R. Humphry, R. Jain, and C.-F. Shu. Virage image search engine: An open framework for image management. In Storage and Retrieval for Image and Video Databases, Proc. SPIE 2670, pages 76 87, [2] T. M. Breuel. Fast recognition using adaptive subdivisions of transformation space. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages , [3] S.-F. Chang, W. Chen, H. J. Meng, H. Sundaram, and D. Zhong. VideoQ: An automated content based video search system using visual cues. In Proceedings of the ACM Multimedia Conference, pages , [4] I. J. Cox, M. L. Miller, T. P. Minka, T. V. Papathomas, and P. N. Yianilos. The Bayesian image retrieval system, PicHunter: Theory, implementation and psychophysical experiments. IEEE Transactions on Image Processing, 9(1):20=37, [5] M. Flickner, H. Sawhney, W. Niblack, J. Ashley, Q.Huang,B.Dom,M.Gorkani,J.Hafner,D.Lee, D. Petkovic, D. Steele, and P. Yanker. The QBIC system. IEEE Computer, 28(9):23 32, Sept [6] T. Gevers and A. W. M. Smeulders. PicToSeek: Combining color and shape invariant features for image retrieval. IEEE Transactions on Image Processing, 9(1): , Jan [7] D. P. Huttenlocher and W. J. Rucklidge. A multi-resolution technique for comparing images using the Hausdorff distance. In Proceedings of the IEEE Recognition, pages , [8] C.-C. Liu and A. L. P. Chen. 3d-List: A data structure for efficient video query processing. IEEE Transactions on Knowledge and Data Engineering, 14(1): , [9] W.-Y. Ma and B. S. Manjunath. NeTra: A toolbox for navigating large image databases. Multimedia Systems, 7(3): , [10] F. Maes, A. Collignon, D. Vandermeulen, G. Marchal, and P. Suetens. Multimodality image registration by maximization of mutual information. IEEE Transactions on Medical Imaging, 16(2): , Apr [11] H. Murase and S. K. Nayar. Visual learning and recognition of 3-d objects from appearance. International Journal of Computer Vision, 14:5 24, [12] C. F. Olson. A probabilistic formulation for Hausdorff matching. In Proceedings of the IEEE Computer Society Recognition, pages , [13] C. F. Olson. Image registration by aligning entropies. In Proceedings of the IEEE Computer Society Recognition, volume 2, pages , [14] C. F. Olson. Maximum-likelihood image matching. IEEE Transactions on Pattern Analysis and Machine Intelligence, 24(6): , June [15] A. Pentland, R. W. Picard, and S. Sclaroff. Photobook: Content-based manipulation of image databases. International Journal of Computer Vision, 18(3): , June [16] W. H. Press, S. A. Teukolsky, W. T. Vetterling, and B. P. Plannery. Numerical Recipes in C. Cambridge University Press, [17] A. Rahimi, L.-P. Morency, and T. Darrell. Reducing drift in parametric motion tracking. In Proceedings of the International Conference on Computer Vision, volume 1, pages , [18] W. J. Rucklidge. Efficient guaranteed search for gray-level patterns. In Proceedings of the IEEE Recognition, pages , [19] M. A. Turk and A. P. Pentland. Face recognition using eigenfaces. In Proceedings of the IEEE Recognition, pages , [20] R. C. Veltkamp and M. Tanase. Content-based image retrieval systems: A survey, revised from technical report UU-CS , Dept. of Computing Science. [21] P. Viola and W. M. Wells. Alignment by maximization of mutual information. International Journal of Computer Vision, 24(2): , 1997.

Fast Image Registration via Joint Gradient Maximization: Application to Multi-Modal Data

Fast Image Registration via Joint Gradient Maximization: Application to Multi-Modal Data MITSUBISHI ELECTRIC RESEARCH LABORATORIES http://www.merl.com Fast Image Registration via Joint Gradient Maximization: Application to Multi-Modal Data Xue Mei, Fatih Porikli TR-19 September Abstract We

More information

Semantic Visual Templates: Linking Visual Features to Semantics

Semantic Visual Templates: Linking Visual Features to Semantics Semantic Visual Templates: Linking Visual Features to Semantics Shih-Fu Chang William Chen Hari Sundaram Dept. of Electrical Engineering Columbia University New York New York 10027. E-mail: fsfchang,bchen,sundaramg@ctr.columbia.edu

More information

Holistic Correlation of Color Models, Color Features and Distance Metrics on Content-Based Image Retrieval

Holistic Correlation of Color Models, Color Features and Distance Metrics on Content-Based Image Retrieval Holistic Correlation of Color Models, Color Features and Distance Metrics on Content-Based Image Retrieval Swapnil Saurav 1, Prajakta Belsare 2, Siddhartha Sarkar 3 1Researcher, Abhidheya Labs and Knowledge

More information

Propagating these values through the probability density function yields a bound on the likelihood score that can be achieved by any position in the c

Propagating these values through the probability density function yields a bound on the likelihood score that can be achieved by any position in the c Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2000. Maximum-Likelihood Template Matching Clark F. Olson Jet Propulsion Laboratory, California Institute of Technology 4800

More information

Image Mining: frameworks and techniques

Image Mining: frameworks and techniques Image Mining: frameworks and techniques Madhumathi.k 1, Dr.Antony Selvadoss Thanamani 2 M.Phil, Department of computer science, NGM College, Pollachi, Coimbatore, India 1 HOD Department of Computer Science,

More information

Maximum-Likelihood Image Matching

Maximum-Likelihood Image Matching IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 24, NO. 6, JUNE 2002 853 Maximum-Likelihood Image Matching Clark F. Olson, Member, IEEE AbstractÐImage matching applications such as

More information

COMPARISON OF SOME CONTENT-BASED IMAGE RETRIEVAL SYSTEMS WITH ROCK TEXTURE IMAGES

COMPARISON OF SOME CONTENT-BASED IMAGE RETRIEVAL SYSTEMS WITH ROCK TEXTURE IMAGES COMPARISON OF SOME CONTENT-BASED IMAGE RETRIEVAL SYSTEMS WITH ROCK TEXTURE IMAGES Leena Lepistö 1, Iivari Kunttu 1, Jorma Autio 2, and Ari Visa 1 1 Tampere University of Technology, Institute of Signal

More information

ELEC Dr Reji Mathew Electrical Engineering UNSW

ELEC Dr Reji Mathew Electrical Engineering UNSW ELEC 4622 Dr Reji Mathew Electrical Engineering UNSW Review of Motion Modelling and Estimation Introduction to Motion Modelling & Estimation Forward Motion Backward Motion Block Motion Estimation Motion

More information

Why study Computer Vision?

Why study Computer Vision? Why study Computer Vision? Images and movies are everywhere Fast-growing collection of useful applications building representations of the 3D world from pictures automated surveillance (who s doing what)

More information

Shape Retrieval with Flat Contour Segments

Shape Retrieval with Flat Contour Segments Shape Retrieval with Flat Contour Segments Dalong Li 1, Steven Simske Intelligent Enterprise Technologies Laboratory HP Laboratories Palo Alto HPL-2002-250 September 9 th, 2002* image database, image retrieval,

More information

Detection and recognition of moving objects using statistical motion detection and Fourier descriptors

Detection and recognition of moving objects using statistical motion detection and Fourier descriptors Detection and recognition of moving objects using statistical motion detection and Fourier descriptors Daniel Toth and Til Aach Institute for Signal Processing, University of Luebeck, Germany toth@isip.uni-luebeck.de

More information

Image Mining Using Image Feature

Image Mining Using Image Feature Image Mining Using Image Feature Varsha Kundlikar 1, Meghana Nagori 2. Abstract In this paper, features of image used for mining images from database. Basic features of images such as colour of pixel in

More information

CS 231A Computer Vision (Fall 2012) Problem Set 3

CS 231A Computer Vision (Fall 2012) Problem Set 3 CS 231A Computer Vision (Fall 2012) Problem Set 3 Due: Nov. 13 th, 2012 (2:15pm) 1 Probabilistic Recursion for Tracking (20 points) In this problem you will derive a method for tracking a point of interest

More information

2 Line-Angle-Ratio Statistics Experiments on various types of images showed us that one of the strongest spatial features of an image is its line segm

2 Line-Angle-Ratio Statistics Experiments on various types of images showed us that one of the strongest spatial features of an image is its line segm Using Texture in Image Similarity and Retrieval Selim Aksoy and Robert M. Haralick Intelligent Systems Laboratory Department of Electrical Engineering University of Washington Seattle, WA 9895-25 faksoy,haralickg@isl.ee.washington.edu

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

Querying by Color Regions using the VisualSEEk Content-Based Visual Query System

Querying by Color Regions using the VisualSEEk Content-Based Visual Query System Querying by Color Regions using the VisualSEEk Content-Based Visual Query System John R. Smith and Shih-Fu Chang Center for Image Technology for New Media and Department of Electrical Engineering Columbia

More information

Wide-Baseline Stereo Vision for Mars Rovers

Wide-Baseline Stereo Vision for Mars Rovers Proceedings of the 2003 IEEE/RSJ Intl. Conference on Intelligent Robots and Systems Las Vegas, Nevada October 2003 Wide-Baseline Stereo Vision for Mars Rovers Clark F. Olson Habib Abi-Rached Ming Ye Jonathan

More information

CBIVR: Content-Based Image and Video Retrieval

CBIVR: Content-Based Image and Video Retrieval CBIVR: Content-Based Image and Video Retrieval QBIC: Query by Image Content QBIC First commercial system Search by: color percentages color layout teture shape/location keywords Try their demo: http://wwwqbic.almaden.ibm.com

More information

Sign Language Recognition using Dynamic Time Warping and Hand Shape Distance Based on Histogram of Oriented Gradient Features

Sign Language Recognition using Dynamic Time Warping and Hand Shape Distance Based on Histogram of Oriented Gradient Features Sign Language Recognition using Dynamic Time Warping and Hand Shape Distance Based on Histogram of Oriented Gradient Features Pat Jangyodsuk Department of Computer Science and Engineering The University

More information

CS 223B Computer Vision Problem Set 3

CS 223B Computer Vision Problem Set 3 CS 223B Computer Vision Problem Set 3 Due: Feb. 22 nd, 2011 1 Probabilistic Recursion for Tracking In this problem you will derive a method for tracking a point of interest through a sequence of images.

More information

Color Local Texture Features Based Face Recognition

Color Local Texture Features Based Face Recognition Color Local Texture Features Based Face Recognition Priyanka V. Bankar Department of Electronics and Communication Engineering SKN Sinhgad College of Engineering, Korti, Pandharpur, Maharashtra, India

More information

An Introduction to Content Based Image Retrieval

An Introduction to Content Based Image Retrieval CHAPTER -1 An Introduction to Content Based Image Retrieval 1.1 Introduction With the advancement in internet and multimedia technologies, a huge amount of multimedia data in the form of audio, video and

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

Feature Descriptors. CS 510 Lecture #21 April 29 th, 2013

Feature Descriptors. CS 510 Lecture #21 April 29 th, 2013 Feature Descriptors CS 510 Lecture #21 April 29 th, 2013 Programming Assignment #4 Due two weeks from today Any questions? How is it going? Where are we? We have two umbrella schemes for object recognition

More information

Binary Histogram in Image Classification for Retrieval Purposes

Binary Histogram in Image Classification for Retrieval Purposes Binary Histogram in Image Classification for Retrieval Purposes Iivari Kunttu 1, Leena Lepistö 1, Juhani Rauhamaa 2, and Ari Visa 1 1 Tampere University of Technology Institute of Signal Processing P.

More information

WATERMARKING FOR LIGHT FIELD RENDERING 1

WATERMARKING FOR LIGHT FIELD RENDERING 1 ATERMARKING FOR LIGHT FIELD RENDERING 1 Alper Koz, Cevahir Çığla and A. Aydın Alatan Department of Electrical and Electronics Engineering, METU Balgat, 06531, Ankara, TURKEY. e-mail: koz@metu.edu.tr, cevahir@eee.metu.edu.tr,

More information

Eppur si muove ( And yet it moves )

Eppur si muove ( And yet it moves ) Eppur si muove ( And yet it moves ) - Galileo Galilei University of Texas at Arlington Tracking of Image Features CSE 4392-5369 Vision-based Robot Sensing, Localization and Control Dr. Gian Luca Mariottini,

More information

Classification. Vladimir Curic. Centre for Image Analysis Swedish University of Agricultural Sciences Uppsala University

Classification. Vladimir Curic. Centre for Image Analysis Swedish University of Agricultural Sciences Uppsala University Classification Vladimir Curic Centre for Image Analysis Swedish University of Agricultural Sciences Uppsala University Outline An overview on classification Basics of classification How to choose appropriate

More information

A Texture Feature Extraction Technique Using 2D-DFT and Hamming Distance

A Texture Feature Extraction Technique Using 2D-DFT and Hamming Distance A Texture Feature Extraction Technique Using 2D-DFT and Hamming Distance Author Tao, Yu, Muthukkumarasamy, Vallipuram, Verma, Brijesh, Blumenstein, Michael Published 2003 Conference Title Fifth International

More information

Sketch Based Image Retrieval Approach Using Gray Level Co-Occurrence Matrix

Sketch Based Image Retrieval Approach Using Gray Level Co-Occurrence Matrix Sketch Based Image Retrieval Approach Using Gray Level Co-Occurrence Matrix K... Nagarjuna Reddy P. Prasanna Kumari JNT University, JNT University, LIET, Himayatsagar, Hyderabad-8, LIET, Himayatsagar,

More information

MULTIVIEW REPRESENTATION OF 3D OBJECTS OF A SCENE USING VIDEO SEQUENCES

MULTIVIEW REPRESENTATION OF 3D OBJECTS OF A SCENE USING VIDEO SEQUENCES MULTIVIEW REPRESENTATION OF 3D OBJECTS OF A SCENE USING VIDEO SEQUENCES Mehran Yazdi and André Zaccarin CVSL, Dept. of Electrical and Computer Engineering, Laval University Ste-Foy, Québec GK 7P4, Canada

More information

CONTENT BASED IMAGE RETRIEVAL SYSTEM USING IMAGE CLASSIFICATION

CONTENT BASED IMAGE RETRIEVAL SYSTEM USING IMAGE CLASSIFICATION International Journal of Research and Reviews in Applied Sciences And Engineering (IJRRASE) Vol 8. No.1 2016 Pp.58-62 gopalax Journals, Singapore available at : www.ijcns.com ISSN: 2231-0061 CONTENT BASED

More information

Automatic Colorization of Grayscale Images

Automatic Colorization of Grayscale Images Automatic Colorization of Grayscale Images Austin Sousa Rasoul Kabirzadeh Patrick Blaes Department of Electrical Engineering, Stanford University 1 Introduction ere exists a wealth of photographic images,

More information

Learning-based Neuroimage Registration

Learning-based Neuroimage Registration Learning-based Neuroimage Registration Leonid Teverovskiy and Yanxi Liu 1 October 2004 CMU-CALD-04-108, CMU-RI-TR-04-59 School of Computer Science Carnegie Mellon University Pittsburgh, PA 15213 Abstract

More information

Segmentation and Tracking of Partial Planar Templates

Segmentation and Tracking of Partial Planar Templates Segmentation and Tracking of Partial Planar Templates Abdelsalam Masoud William Hoff Colorado School of Mines Colorado School of Mines Golden, CO 800 Golden, CO 800 amasoud@mines.edu whoff@mines.edu Abstract

More information

CS 543: Final Project Report Texture Classification using 2-D Noncausal HMMs

CS 543: Final Project Report Texture Classification using 2-D Noncausal HMMs CS 543: Final Project Report Texture Classification using 2-D Noncausal HMMs Felix Wang fywang2 John Wieting wieting2 Introduction We implement a texture classification algorithm using 2-D Noncausal Hidden

More information

Human Detection. A state-of-the-art survey. Mohammad Dorgham. University of Hamburg

Human Detection. A state-of-the-art survey. Mohammad Dorgham. University of Hamburg Human Detection A state-of-the-art survey Mohammad Dorgham University of Hamburg Presentation outline Motivation Applications Overview of approaches (categorized) Approaches details References Motivation

More information

Textural Features for Image Database Retrieval

Textural Features for Image Database Retrieval Textural Features for Image Database Retrieval Selim Aksoy and Robert M. Haralick Intelligent Systems Laboratory Department of Electrical Engineering University of Washington Seattle, WA 98195-2500 {aksoy,haralick}@@isl.ee.washington.edu

More information

2D-3D Registration using Gradient-based MI for Image Guided Surgery Systems

2D-3D Registration using Gradient-based MI for Image Guided Surgery Systems 2D-3D Registration using Gradient-based MI for Image Guided Surgery Systems Yeny Yim 1*, Xuanyi Chen 1, Mike Wakid 1, Steve Bielamowicz 2, James Hahn 1 1 Department of Computer Science, The George Washington

More information

Eye Detection by Haar wavelets and cascaded Support Vector Machine

Eye Detection by Haar wavelets and cascaded Support Vector Machine Eye Detection by Haar wavelets and cascaded Support Vector Machine Vishal Agrawal B.Tech 4th Year Guide: Simant Dubey / Amitabha Mukherjee Dept of Computer Science and Engineering IIT Kanpur - 208 016

More information

Research Article Image Retrieval using Clustering Techniques. K.S.Rangasamy College of Technology,,India. K.S.Rangasamy College of Technology, India.

Research Article Image Retrieval using Clustering Techniques. K.S.Rangasamy College of Technology,,India. K.S.Rangasamy College of Technology, India. Journal of Recent Research in Engineering and Technology 3(1), 2016, pp21-28 Article ID J11603 ISSN (Online): 2349 2252, ISSN (Print):2349 2260 Bonfay Publications, 2016 Research Article Image Retrieval

More information

Chapter 9 Object Tracking an Overview

Chapter 9 Object Tracking an Overview Chapter 9 Object Tracking an Overview The output of the background subtraction algorithm, described in the previous chapter, is a classification (segmentation) of pixels into foreground pixels (those belonging

More information

Welcome Back to Fundamental of Multimedia (MR412) Fall, ZHU Yongxin, Winson

Welcome Back to Fundamental of Multimedia (MR412) Fall, ZHU Yongxin, Winson Welcome Back to Fundamental of Multimedia (MR412) Fall, 2012 ZHU Yongxin, Winson zhuyongxin@sjtu.edu.cn Content-Based Retrieval in Digital Libraries 18.1 How Should We Retrieve Images? 18.2 C-BIRD : A

More information

BAG-OF-VISUAL WORDS (BoVW) MODEL BASED APPROACH FOR CONTENT BASED IMAGE RETRIEVAL (CBIR) IN PEER TO PEER (P2P)NETWORKS.

BAG-OF-VISUAL WORDS (BoVW) MODEL BASED APPROACH FOR CONTENT BASED IMAGE RETRIEVAL (CBIR) IN PEER TO PEER (P2P)NETWORKS. BAG-OF-VISUAL WORDS (BoVW) MODEL BASED APPROACH FOR CONTENT BASED IMAGE RETRIEVAL (CBIR) IN PEER TO PEER (P2P)NETWORKS. 1 R.Lavanya, 2 E.Lavanya, 1 PG Scholar, Dept Of Computer Science Engineering,Mailam

More information

Recognition (Part 4) Introduction to Computer Vision CSE 152 Lecture 17

Recognition (Part 4) Introduction to Computer Vision CSE 152 Lecture 17 Recognition (Part 4) CSE 152 Lecture 17 Announcements Homework 5 is due June 9, 11:59 PM Reading: Chapter 15: Learning to Classify Chapter 16: Classifying Images Chapter 17: Detecting Objects in Images

More information

Local Features Tutorial: Nov. 8, 04

Local Features Tutorial: Nov. 8, 04 Local Features Tutorial: Nov. 8, 04 Local Features Tutorial References: Matlab SIFT tutorial (from course webpage) Lowe, David G. Distinctive Image Features from Scale Invariant Features, International

More information

1 Introduction Motivation and Aims Functional Imaging Computational Neuroanatomy... 12

1 Introduction Motivation and Aims Functional Imaging Computational Neuroanatomy... 12 Contents 1 Introduction 10 1.1 Motivation and Aims....... 10 1.1.1 Functional Imaging.... 10 1.1.2 Computational Neuroanatomy... 12 1.2 Overview of Chapters... 14 2 Rigid Body Registration 18 2.1 Introduction.....

More information

Digital Image Processing

Digital Image Processing Digital Image Processing Part 9: Representation and Description AASS Learning Systems Lab, Dep. Teknik Room T1209 (Fr, 11-12 o'clock) achim.lilienthal@oru.se Course Book Chapter 11 2011-05-17 Contents

More information

Colorado School of Mines. Computer Vision. Professor William Hoff Dept of Electrical Engineering &Computer Science.

Colorado School of Mines. Computer Vision. Professor William Hoff Dept of Electrical Engineering &Computer Science. Professor William Hoff Dept of Electrical Engineering &Computer Science http://inside.mines.edu/~whoff/ 1 Statistical Models for Shape and Appearance Note some material for these slides came from Algorithms

More information

Object Category Detection: Sliding Windows

Object Category Detection: Sliding Windows 04/10/12 Object Category Detection: Sliding Windows Computer Vision CS 543 / ECE 549 University of Illinois Derek Hoiem Today s class: Object Category Detection Overview of object category detection Statistical

More information

CS 664 Segmentation. Daniel Huttenlocher

CS 664 Segmentation. Daniel Huttenlocher CS 664 Segmentation Daniel Huttenlocher Grouping Perceptual Organization Structural relationships between tokens Parallelism, symmetry, alignment Similarity of token properties Often strong psychophysical

More information

Using texture to analyze and manage large collections of remote sensed image and video data

Using texture to analyze and manage large collections of remote sensed image and video data Using texture to analyze and manage large collections of remote sensed image and video data Shawn Newsam, Lei Wang, Sitaram Bhagavathy, and Bangalore S. Manjunath We describe recent research into using

More information

Rotation Invariant Image Registration using Robust Shape Matching

Rotation Invariant Image Registration using Robust Shape Matching International Journal of Electronic and Electrical Engineering. ISSN 0974-2174 Volume 7, Number 2 (2014), pp. 125-132 International Research Publication House http://www.irphouse.com Rotation Invariant

More information

A Miniature-Based Image Retrieval System

A Miniature-Based Image Retrieval System A Miniature-Based Image Retrieval System Md. Saiful Islam 1 and Md. Haider Ali 2 Institute of Information Technology 1, Dept. of Computer Science and Engineering 2, University of Dhaka 1, 2, Dhaka-1000,

More information

An Approach for Reduction of Rain Streaks from a Single Image

An Approach for Reduction of Rain Streaks from a Single Image An Approach for Reduction of Rain Streaks from a Single Image Vijayakumar Majjagi 1, Netravati U M 2 1 4 th Semester, M. Tech, Digital Electronics, Department of Electronics and Communication G M Institute

More information

Depth. Common Classification Tasks. Example: AlexNet. Another Example: Inception. Another Example: Inception. Depth

Depth. Common Classification Tasks. Example: AlexNet. Another Example: Inception. Another Example: Inception. Depth Common Classification Tasks Recognition of individual objects/faces Analyze object-specific features (e.g., key points) Train with images from different viewing angles Recognition of object classes Analyze

More information

SIFT - scale-invariant feature transform Konrad Schindler

SIFT - scale-invariant feature transform Konrad Schindler SIFT - scale-invariant feature transform Konrad Schindler Institute of Geodesy and Photogrammetry Invariant interest points Goal match points between images with very different scale, orientation, projective

More information

A Novel Image Semantic Understanding and Feature Extraction Algorithm. and Wenzhun Huang

A Novel Image Semantic Understanding and Feature Extraction Algorithm. and Wenzhun Huang A Novel Image Semantic Understanding and Feature Extraction Algorithm Xinxin Xie 1, a 1, b* and Wenzhun Huang 1 School of Information Engineering, Xijing University, Xi an 710123, China a 346148500@qq.com,

More information

Content Based Image Retrieval (CBIR) Using Segmentation Process

Content Based Image Retrieval (CBIR) Using Segmentation Process Content Based Image Retrieval (CBIR) Using Segmentation Process R.Gnanaraja 1, B. Jagadishkumar 2, S.T. Premkumar 3, B. Sunil kumar 4 1, 2, 3, 4 PG Scholar, Department of Computer Science and Engineering,

More information

Announcements. Recognition (Part 3) Model-Based Vision. A Rough Recognition Spectrum. Pose consistency. Recognition by Hypothesize and Test

Announcements. Recognition (Part 3) Model-Based Vision. A Rough Recognition Spectrum. Pose consistency. Recognition by Hypothesize and Test Announcements (Part 3) CSE 152 Lecture 16 Homework 3 is due today, 11:59 PM Homework 4 will be assigned today Due Sat, Jun 4, 11:59 PM Reading: Chapter 15: Learning to Classify Chapter 16: Classifying

More information

Content Based Image Retrieval with Semantic Features using Object Ontology

Content Based Image Retrieval with Semantic Features using Object Ontology Content Based Image Retrieval with Semantic Features using Object Ontology Anuja Khodaskar Research Scholar College of Engineering & Technology, Amravati, India Dr. S.A. Ladke Principal Sipna s College

More information

Random projection for non-gaussian mixture models

Random projection for non-gaussian mixture models Random projection for non-gaussian mixture models Győző Gidófalvi Department of Computer Science and Engineering University of California, San Diego La Jolla, CA 92037 gyozo@cs.ucsd.edu Abstract Recently,

More information

Object Detection by Point Feature Matching using Matlab

Object Detection by Point Feature Matching using Matlab Object Detection by Point Feature Matching using Matlab 1 Faishal Badsha, 2 Rafiqul Islam, 3,* Mohammad Farhad Bulbul 1 Department of Mathematics and Statistics, Bangladesh University of Business and Technology,

More information

Content based Image Retrievals for Brain Related Diseases

Content based Image Retrievals for Brain Related Diseases Content based Image Retrievals for Brain Related Diseases T.V. Madhusudhana Rao Department of CSE, T.P.I.S.T., Bobbili, Andhra Pradesh, INDIA S. Pallam Setty Department of CS&SE, Andhra University, Visakhapatnam,

More information

Image Segmentation and Registration

Image Segmentation and Registration Image Segmentation and Registration Dr. Christine Tanner (tanner@vision.ee.ethz.ch) Computer Vision Laboratory, ETH Zürich Dr. Verena Kaynig, Machine Learning Laboratory, ETH Zürich Outline Segmentation

More information

COMPUTER AND ROBOT VISION

COMPUTER AND ROBOT VISION VOLUME COMPUTER AND ROBOT VISION Robert M. Haralick University of Washington Linda G. Shapiro University of Washington T V ADDISON-WESLEY PUBLISHING COMPANY Reading, Massachusetts Menlo Park, California

More information

A Texture-based Method for Detecting Moving Objects

A Texture-based Method for Detecting Moving Objects A Texture-based Method for Detecting Moving Objects M. Heikkilä, M. Pietikäinen and J. Heikkilä Machine Vision Group Infotech Oulu and Department of Electrical and Information Engineering P.O. Box 4500

More information

CSI 4107 Image Information Retrieval

CSI 4107 Image Information Retrieval CSI 4107 Image Information Retrieval This slides are inspired by a tutorial on Medical Image Retrieval by Henning Müller and Thomas Deselaers, 2005-2006 1 Outline Introduction Content-based image retrieval

More information

Robotics Programming Laboratory

Robotics Programming Laboratory Chair of Software Engineering Robotics Programming Laboratory Bertrand Meyer Jiwon Shin Lecture 8: Robot Perception Perception http://pascallin.ecs.soton.ac.uk/challenges/voc/databases.html#caltech car

More information

Probabilistic Facial Feature Extraction Using Joint Distribution of Location and Texture Information

Probabilistic Facial Feature Extraction Using Joint Distribution of Location and Texture Information Probabilistic Facial Feature Extraction Using Joint Distribution of Location and Texture Information Mustafa Berkay Yilmaz, Hakan Erdogan, Mustafa Unel Sabanci University, Faculty of Engineering and Natural

More information

Journal of Asian Scientific Research FEATURES COMPOSITION FOR PROFICIENT AND REAL TIME RETRIEVAL IN CBIR SYSTEM. Tohid Sedghi

Journal of Asian Scientific Research FEATURES COMPOSITION FOR PROFICIENT AND REAL TIME RETRIEVAL IN CBIR SYSTEM. Tohid Sedghi Journal of Asian Scientific Research, 013, 3(1):68-74 Journal of Asian Scientific Research journal homepage: http://aessweb.com/journal-detail.php?id=5003 FEATURES COMPOSTON FOR PROFCENT AND REAL TME RETREVAL

More information

Announcements. Computer Vision I. Motion Field Equation. Revisiting the small motion assumption. Visual Tracking. CSE252A Lecture 19.

Announcements. Computer Vision I. Motion Field Equation. Revisiting the small motion assumption. Visual Tracking. CSE252A Lecture 19. Visual Tracking CSE252A Lecture 19 Hw 4 assigned Announcements No class on Thursday 12/6 Extra class on Tuesday 12/4 at 6:30PM in WLH Room 2112 Motion Field Equation Measurements I x = I x, T: Components

More information

Occlusion Robust Multi-Camera Face Tracking

Occlusion Robust Multi-Camera Face Tracking Occlusion Robust Multi-Camera Face Tracking Josh Harguess, Changbo Hu, J. K. Aggarwal Computer & Vision Research Center / Department of ECE The University of Texas at Austin harguess@utexas.edu, changbo.hu@gmail.com,

More information

Percentile Blobs for Image Similarity

Percentile Blobs for Image Similarity Percentile Blobs for Image Similarity Nicholas R. Howe Cornell University Department of Computer Science Ithaca, NY 14853 nihowe@cs.cornell.edu Abstract We present a new algorithm called PBSIM for computing

More information

Pictures at an Exhibition

Pictures at an Exhibition Pictures at an Exhibition Han-I Su Department of Electrical Engineering Stanford University, CA, 94305 Abstract We employ an image identification algorithm for interactive museum guide with pictures taken

More information

Classifying Images with Visual/Textual Cues. By Steven Kappes and Yan Cao

Classifying Images with Visual/Textual Cues. By Steven Kappes and Yan Cao Classifying Images with Visual/Textual Cues By Steven Kappes and Yan Cao Motivation Image search Building large sets of classified images Robotics Background Object recognition is unsolved Deformable shaped

More information

EE 701 ROBOT VISION. Segmentation

EE 701 ROBOT VISION. Segmentation EE 701 ROBOT VISION Regions and Image Segmentation Histogram-based Segmentation Automatic Thresholding K-means Clustering Spatial Coherence Merging and Splitting Graph Theoretic Segmentation Region Growing

More information

Feature descriptors and matching

Feature descriptors and matching Feature descriptors and matching Detections at multiple scales Invariance of MOPS Intensity Scale Rotation Color and Lighting Out-of-plane rotation Out-of-plane rotation Better representation than color:

More information

PERFORMANCE EVALUATION OF ONTOLOGY AND FUZZYBASE CBIR

PERFORMANCE EVALUATION OF ONTOLOGY AND FUZZYBASE CBIR PERFORMANCE EVALUATION OF ONTOLOGY AND FUZZYBASE CBIR ABSTRACT Tajman sandhu (Research scholar) Department of Information Technology Chandigarh Engineering College, Landran, Punjab, India yuvi_taj@yahoo.com

More information

Texture Similarity Measure. Pavel Vácha. Institute of Information Theory and Automation, AS CR Faculty of Mathematics and Physics, Charles University

Texture Similarity Measure. Pavel Vácha. Institute of Information Theory and Automation, AS CR Faculty of Mathematics and Physics, Charles University Texture Similarity Measure Pavel Vácha Institute of Information Theory and Automation, AS CR Faculty of Mathematics and Physics, Charles University What is texture similarity? Outline 1. Introduction Julesz

More information

Elliptical Head Tracker using Intensity Gradients and Texture Histograms

Elliptical Head Tracker using Intensity Gradients and Texture Histograms Elliptical Head Tracker using Intensity Gradients and Texture Histograms Sriram Rangarajan, Dept. of Electrical and Computer Engineering, Clemson University, Clemson, SC 29634 srangar@clemson.edu December

More information

Object Detection System

Object Detection System A Trainable View-Based Object Detection System Thesis Proposal Henry A. Rowley Thesis Committee: Takeo Kanade, Chair Shumeet Baluja Dean Pomerleau Manuela Veloso Tomaso Poggio, MIT Motivation Object detection

More information

Tri-modal Human Body Segmentation

Tri-modal Human Body Segmentation Tri-modal Human Body Segmentation Master of Science Thesis Cristina Palmero Cantariño Advisor: Sergio Escalera Guerrero February 6, 2014 Outline 1 Introduction 2 Tri-modal dataset 3 Proposed baseline 4

More information

A Comparative Analysis of Retrieval Techniques in Content Based Image Retrieval

A Comparative Analysis of Retrieval Techniques in Content Based Image Retrieval A Comparative Analysis of Retrieval Techniques in Content Based Image Retrieval Mohini. P. Sardey 1, G. K. Kharate 2 1 AISSMS Institute Of Information Technology, Savitribai Phule Pune University, Pune

More information

Learning to Recognize Faces in Realistic Conditions

Learning to Recognize Faces in Realistic Conditions 000 001 002 003 004 005 006 007 008 009 010 011 012 013 014 015 016 017 018 019 020 021 022 023 024 025 026 027 028 029 030 031 032 033 034 035 036 037 038 039 040 041 042 043 044 045 046 047 048 049 050

More information

Hierarchical Building Recognition

Hierarchical Building Recognition Hierarchical Building Recognition Anonymous Abstract We propose a novel and efficient method for recognition of buildings and locations with dominant man-made structures. The method exploits two different

More information

SIFT: SCALE INVARIANT FEATURE TRANSFORM SURF: SPEEDED UP ROBUST FEATURES BASHAR ALSADIK EOS DEPT. TOPMAP M13 3D GEOINFORMATION FROM IMAGES 2014

SIFT: SCALE INVARIANT FEATURE TRANSFORM SURF: SPEEDED UP ROBUST FEATURES BASHAR ALSADIK EOS DEPT. TOPMAP M13 3D GEOINFORMATION FROM IMAGES 2014 SIFT: SCALE INVARIANT FEATURE TRANSFORM SURF: SPEEDED UP ROBUST FEATURES BASHAR ALSADIK EOS DEPT. TOPMAP M13 3D GEOINFORMATION FROM IMAGES 2014 SIFT SIFT: Scale Invariant Feature Transform; transform image

More information

Video Alignment. Literature Survey. Spring 2005 Prof. Brian Evans Multidimensional Digital Signal Processing Project The University of Texas at Austin

Video Alignment. Literature Survey. Spring 2005 Prof. Brian Evans Multidimensional Digital Signal Processing Project The University of Texas at Austin Literature Survey Spring 2005 Prof. Brian Evans Multidimensional Digital Signal Processing Project The University of Texas at Austin Omer Shakil Abstract This literature survey compares various methods

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

[2006] IEEE. Reprinted, with permission, from [Wenjing Jia, Huaifeng Zhang, Xiangjian He, and Qiang Wu, A Comparison on Histogram Based Image

[2006] IEEE. Reprinted, with permission, from [Wenjing Jia, Huaifeng Zhang, Xiangjian He, and Qiang Wu, A Comparison on Histogram Based Image [6] IEEE. Reprinted, with permission, from [Wenjing Jia, Huaifeng Zhang, Xiangjian He, and Qiang Wu, A Comparison on Histogram Based Image Matching Methods, Video and Signal Based Surveillance, 6. AVSS

More information

Abstract. 1 Introduction. 2 Motivation. Information and Communication Engineering October 29th 2010

Abstract. 1 Introduction. 2 Motivation. Information and Communication Engineering October 29th 2010 Information and Communication Engineering October 29th 2010 A Survey on Head Pose Estimation from Low Resolution Image Sato Laboratory M1, 48-106416, Isarun CHAMVEHA Abstract Recognizing the head pose

More information

A Survey on Content Based Image Retrieval

A Survey on Content Based Image Retrieval A Survey on Content Based Image Retrieval Aniket Mirji 1, Danish Sudan 2, Rushabh Kagwade 3, Savita Lohiya 4 U.G. Students of Department of Information Technology, SIES GST, Mumbai, Maharashtra, India

More information

Multi-modal Image Registration Using the Generalized Survival Exponential Entropy

Multi-modal Image Registration Using the Generalized Survival Exponential Entropy Multi-modal Image Registration Using the Generalized Survival Exponential Entropy Shu Liao and Albert C.S. Chung Lo Kwee-Seong Medical Image Analysis Laboratory, Department of Computer Science and Engineering,

More information

2 Proposed Methodology

2 Proposed Methodology 3rd International Conference on Multimedia Technology(ICMT 2013) Object Detection in Image with Complex Background Dong Li, Yali Li, Fei He, Shengjin Wang 1 State Key Laboratory of Intelligent Technology

More information

A Novel Image Super-resolution Reconstruction Algorithm based on Modified Sparse Representation

A Novel Image Super-resolution Reconstruction Algorithm based on Modified Sparse Representation , pp.162-167 http://dx.doi.org/10.14257/astl.2016.138.33 A Novel Image Super-resolution Reconstruction Algorithm based on Modified Sparse Representation Liqiang Hu, Chaofeng He Shijiazhuang Tiedao University,

More information

A Content Based Image Retrieval System Based on Color Features

A Content Based Image Retrieval System Based on Color Features A Content Based Image Retrieval System Based on Features Irena Valova, University of Rousse Angel Kanchev, Department of Computer Systems and Technologies, Rousse, Bulgaria, Irena@ecs.ru.acad.bg Boris

More information

William Yang Group 14 Mentor: Dr. Rogerio Richa Visual Tracking of Surgical Tools in Retinal Surgery using Particle Filtering

William Yang Group 14 Mentor: Dr. Rogerio Richa Visual Tracking of Surgical Tools in Retinal Surgery using Particle Filtering Mutual Information Computation and Maximization Using GPU Yuping Lin and Gérard Medioni Computer Vision and Pattern Recognition Workshops (CVPR) Anchorage, AK, pp. 1-6, June 2008 Project Summary and Paper

More information

Robust Realignment of fmri Time Series Data

Robust Realignment of fmri Time Series Data Robust Realignment of fmri Time Series Data Ben Dodson bjdodson@stanford.edu Olafur Gudmundsson olafurg@stanford.edu December 12, 2008 Abstract FMRI data has become an increasingly popular source for exploring

More information

Blobworld: A System for Region-Based Image Indexing and Retrieval

Blobworld: A System for Region-Based Image Indexing and Retrieval Blobworld: A System for Region-Based Image Indexing and Retrieval Chad Carson, Megan Thomas, Serge Belongie, Joseph M. Hellerstein, and Jitendra Malik EECS Department University of California, Berkeley,

More information

Face Recognition Technology Based On Image Processing Chen Xin, Yajuan Li, Zhimin Tian

Face Recognition Technology Based On Image Processing Chen Xin, Yajuan Li, Zhimin Tian 4th International Conference on Machinery, Materials and Computing Technology (ICMMCT 2016) Face Recognition Technology Based On Image Processing Chen Xin, Yajuan Li, Zhimin Tian Hebei Engineering and

More information