Hybrid ART/Kohonen Neural Model for Document Image Compression. Computer Science Department NMT Socorro, New Mexico {hss,

Size: px
Start display at page:

Download "Hybrid ART/Kohonen Neural Model for Document Image Compression. Computer Science Department NMT Socorro, New Mexico {hss,"

Transcription

1 Hybrid ART/Kohonen Neural Model for Document Image Compression Hamdy S. Soliman Mohammed Omari Computer Science Department NMT Socorro, New Mexico {hss, Keyword -1: Learning Algorithms and Training. Keyword -2: Adaptive Learning, Algorithms Network Resonance, ART, Applications Neura Networks, Comparisons of Networks, Data Compression, Experimental Design, Hybrid Artificial Neural, Image Compression, Kohonen Networks, Minimum Distance Classifier, Neural Network Algorithms/Applications/Architecture, Neural Network Training, Self-Adaptive Training, Self Organizing Neural Networks, Signal to Noise Ratio, Testing Importance, Training Vector Quantization, Unsupervised Classification/Learning, Vector Quantization, Wavelets, Winner-Take-All. Abstract A new neural direct classification model, DC, is introduced for image data compression. It is based on the Adaptive Resonance Theorem (ART1) and the Kohonen Self Organizing Feature Map (KSOFM) neural models. The input subimage domain is clustered into classes of similar subimages. Each class will have a center of mass subimage representative, called centroid. The main goal of the DC training/compression phase is to develop a table of all class' centroids (codebook) and map each input subimage into its class centroid index in the codebook. The process of deciding the membership of an input subimage to some hosting class is similar to the ART1 mechanism. However, the updating of the host class centroid is related to the KSOFM

2 approach. An input image is compressed into a set of centroid' indices, one per subimage, and a local codebook. Compression is realized because the bit size of the input subimage is an order of magnitude larger than its encoding centroid index. The DC model has achieved much better results than the state-of-the-art peer document compression techniques, e.g., JPEG, DjVu, Wavelet, etc. I. Introduction Uncompressed multimedia (graphics, audio and video) data require considerable storage capacity and transmission bandwidth. Thus, image compression is a very important factor for better bandwidth utilization over a communication network. The compression process is based on redundancy and irrelevancy reduction, which are associated with the image domain. Redundancy reduction aims at removing duplication from the signal source (image/video), whereas irrelevancy reduction omits parts of the signal that will not be noticed by the signal receiver, namely the Human Visual System (HVS). There are two well-known major approaches to the implementation of image compression: discrete cosine transform technique (e.g., Wavelet. JPEG, DjVu, ), and neural networks models (e.g., counterpropagation, Kohonen, backpropagation, adaptive resonance theory, ). In the discrete cosine transform approach, compression is accomplished by applying a linear transform to de-correlate the image data (source encoder), quantizing the resulting transform coefficients (quantizer), and entropy-coding the quantized values (entropy encoder). In the neural network approach, a set of input subimage vectors is presented to the net in order to train its neuron's set of weight vectors (weight matrix) via a process of weight adjustment. The

3 neural network approach encompassing the use of the counterpropagation network (CPN) and the Kohonen feature map (KSOFM) models to implement data compression was first suggested by Kohonen and Hecht-Nielson [Kohonen 1982; Kohonen 1984; Hecht-Nielson 1987; Hecht- Nielson 1988]. We introduced a new neural model of direct classification (DC) that is based on the philosophy of ART (adaptive resonance theory) and the KSOFM. Our DC model combines the advantages of these two neural net models in order to achieve high (decompressed) image quality at high compression ratio. The DC model performed exceptionally well in the image document domain, compared to the peer "state-of-art" models (e.g., Wavelet and DjVu). In section II, we briefly introduce to the above neural models. Our DC model is defined in Section III. In Section IV, some experiments, using the traditional peer models versus our DC model, and a discussion of the significant of the obtained results will be covered. Section V will contain our conclusion and future work. II. Utilized Neural Networks Models: The AVQ theory is one of the most recent techniques used in the domain of image compression [1][2][3][7][11][12]. In the AVQ image compression-based approach, the input image is divided into equal size sections (subimages), where each section of N 2 pixels is considered as a vector in the encoding space R HxW. The neural net clusters the subimages into classes of similar subimages. Each class has a representative vector, the centroid, that represents

4 any of its member subimages. All centroid representatives for an image are to be tabled in a lookup table or codebook. Then, each subimage is compressed into the index of its corresponding class representative in the codebook. Thus, the compressed image is a set of representative indices that represent (in order) its original subimages. The compression is realized because the size of the original subimage (N 2 ) is an order of magnitude larger than its corresponding centroid index (log 2 codebook ). The Kohonen self-organizing feature map (KSOFM) is a structured network with a twodimensional array of nodes in which an adaptive partner weight vector is associated with every node [1][2][3]. Based on the AVQ theory, the KSOFM initializes the weight matrix randomly. The input subimage vectors are presented sequentially to the network [9][10]. During the training process, the nearest weight vector (winner) to the input vector is adaptively moved closer to such input vector. Consequently, with every introduction of an input subimage vector, the winning vector moves toward some theoretical class's center of mass (centroid) in its way to become the true representative of all subimages that have won and so similar. The input vector is associated to the class of its corresponding representative centroid. The input subimage vectors are presented in epochs until no change of memberships is observed. The adaptive resonance theory 1 (ART1) was developed by Carpenter and Grossberg (1987, 1988) [4][5][6]. It serves the purpose of cluster discovery. Like networks using a single Kohonen layer with competitive learning neurons, this network learns clusters in an unsupervised mode. The novel property of the ART1 network is the controlled discovery of clusters. In addition, the ART1 network can accommodate new clusters without affecting the storage or recall capabilities for clusters already learned (elasticity/plasticity property).

5 Essentially, the network follows the leader after it originates the first cluster with the first input pattern received. It then creates the second cluster if the distance of the second pattern exceeds a certain threshold; otherwise the pattern is clustered with the first cluster. These features of elasticity and single (one epoch) training cycle truly characterize the ART model. III. Direct Classification Model (DC): Based on AVQ theory, we designed our new Direct Classification neural net engine for image compression/decompression. It follows the winner-take-all feature of the Kohonen model, and the elasticity along with the single epoch training cycle features of the ART1 model. The advantage of the DC over Kohonen is that the input domain is presented only once to the DC system. Therefore the asymptotic training time complexity is O(n), where n is the size of the input domain of subimages; a huge reduction to the original KSOFM time complexity of n 2.The traditional ART1 generates small real numbers in the weight matrix, which might cause the deviation of the classification, in case of low system precision. Instead, our DC model develops classes' representatives (centers of mass) called centroids, which are of the same type of the input subimages (integer vectors). Moreover, both of the Kohonen and ART1 models use training input sets of much smaller size than the total input domain. Therefore, the codebook will be only a good representative of the input training subimages, not the total subimages domain. Such limitation does not exist in our DC model. Our work is in the same area as the well known Wavelet and JPEG image compression techniques, with a major difference in the formation of the lookup tables. The manufacturing of

6 the DC's lookup tables is carried out via the training of a hybrid neural model of the SOFM and ART nets, with some modifications. Our model is a vector quantizer (VQ), which encodes subimage vectors via the mapping of many similar k-dimensional input vectors (with respect to a given distortion measure) into one representative codeword (centroid) vector. The similarity measure of vectors X and Y is based on the distance X-Y (distortion) between X and Y. A collection of codewords is to be stored in a lookup table (codebook), which is utilized later at the decompression phase, in the lookup process. The manufacturing and the nature of the codebook are the key distinctions of our work from other peer mechanisms. The training subphase (Fig. 1) starts by dividing the total subimages vector domain into classes (clusters) of similar subimages. Each class has a centroid that represents the center of mass (average) of all of its member subimages, and its index will be used to replace each of them in the compressed image file. As shown in Fig.2, the mechanism incrementally constructs a codebook of centroids to be placed in the DC neurons synaptic weight vectors. Starting with a set of empty synaptic weights (empty code book), a very simplified ART approach is adopted to build up the codebook table. For each input subimage S, we adaptively train the DC synaptic weights based on the SOFM "winner takes all" concept. The first step in the process of adjustment is to find the closest centroids to S, in the active entries of the codebook, and to obtain a set of possible winning centroids, PWCS. This is done based on an acceptable number of pixel differences (within MaxIntDiff threshold) between S and the codebook entries. Failing to find a PWS (i.e., empty PWS) forces the system to add a new cluster with S as its centroid, given that there is an available entry in the codebook. If not, the system will consider the whole codebook as S's PWCS. Then, among the obtained PWCS, a second

7 process is initiated to find the final wining centroid, WC, for S. WC should have the smallest mean error (ME) with S, among all members of the PWCS. In order to maintain the adaptation feature of the KSOFM model, S is used to adjust WC in order to represent the average of its class members, including S. Another DC variation from the Kohonen model is that only a subset of the class members will be involved in the adjustment of the centroid. The DC philosophy is that only the first k-classified subimages adjust the centroid. Such approach helps in the stability of the centroid ability of accurate representation of the class core, based in the spatial locality. After adjusting WC with the k-spatially consecutive input subimages, the k+1 th input subimage member will have two possibilities. First, it is very similar to the centroid. Therefore not involving it in the centroid updating will not hurt the process. On the other hand, if it is not similar, we do not want it to affect the true centroid representation of the previously classified k- subimages. Thus, the process of adjusting WC is performed only when its class size is still below a certain threshold (MaxTrainSet). The compression subphase only assigns the WC index in the codebook table as the S s compression index (CI). The indices will be stored, in the same sequence as the introduction of the training subimages. The result of the previous operations (subphases) is a codebook table of a fully trained neuron synaptic weight vectors (centroids), and an indices sequence that represents the compression of the original subimages of the input image file, with each subimage denoted by its corresponding centroid representative. The codebook is to be appended to the indices sequence, forming the final compressed file. The compression is achieved due to the huge reduction (order of magnitude) of the subimage representation.

8 In the decompression phase (Fig. 3), the indices sequence in the compressed file is scanned sequentially. For each index value i, at location j, a lookup process is performed in the codebook to obtain the i th centroid that represents the j th original subimage in the decompressed file. In order to achieve a better compression ratio, the lossless LZW is performed on the compressed file before storing (or transmitting) it. Then, at the decompression phase, LZW decompression is performed on the indices file (Fig. 4). VI. Results and discussion We experimented with wide varieties of different images. We drew intensity graphs and pixel value distributions in order to find a balance of the training parameters that achieve good performance (L: codebook size, MaxIntDif: maximum acceptable byte by byte intensity difference between a subimage and a selected centroid, MaxTrainSet: the maximum number of inputs allowed to adjust a specific centroid). A simple conclusion seems very difficult to achieve after a considerable number of experiments using our DC model. The lack of common characteristics among images, even of the same domain, led us to try different parameter values for each image. Although DjVu s developers claim that it treats the background of the image (especially documents) separately [8], experiments showed that DC is better at background recognition (Fig. 6). The DC recognizes the background as the centroids representatives of the highest prominent subimages. Placing such centroids near the start of the codebook (index 0, 1, 2, ) will guarantee many zeros in their binary representation. This will lead the indices files to have many runs of zeros since those centroids indices are the most repeated inside the compressed file.

9 Next table shows the cases, where good results were obtained. Thus, using LZW will result in significant improvement in compression ratio. In Table 1, DC achieves better compression ratio than DjVu. DjVu failed to surpass the DC model in the domain of non-purely-textual documents (mostly graphs and figures). In some examples, as in cdoc12.ppm (Fig. 7), the compression ratio was nearly three times that of the DjVu, with a lossless restoration of the original file. Such excellent results are achieved since the number of different subimage patterns inside the original image is small. The ability to tune the quality and the compression ratio is maintained in the DC model. Such flexibility is achievable due to the private generation of the codebook for each image. Such a feature is not amenable to the universal codebook that is generated for multiple images [1]. V. Conclusion and Future Work The DC hybrid approach of the two prominent neural models, Kohonen (KSOFM) and ART, led to a very useful and practical image compression model. It combines the best features of both, fast training (ART) and reliable/accurate centroid formation (KSOFM). Moreover, our DC model adds its own special features (e.g., introduction of different thresholds to control the centroid development), which allowed it to compete with other well-known peer techniques (e.g., Wavelet, DjVue, JPEG) in the document compression domain. Our future work will involve the expansion of the DC model into the other image domains (e.g., pictorial, satellite, landscape, etc).

10 References [1] H.S. Soliman, A. Abdelali. Toward lossless image compression. Proceedings of the ISCA 8 th International Conference, Denver, CO June 24-26, [2] Tuevo Kohonen. Self-organizing maps, optimization approaches. Artificial Neural Networks, p , Elsevier Science Publishers B.V. (North-Holland) [3] Willie Chang. Neural Networks as a vector quantization image coding model. NMT PHD dissertation, 1993 [4] Jack M.Zurada. Introduction to Artificial Neural Systems. PWS Publishing Company, [5] Naser M. Nasrabadi, Aggelos K.Katsaggelos. Applications of Artificial Neural Network in Image Processing III. San Jose, California, April [6] Naser M. Nasrabadi, Aggelos K.Katsaggelos. Applications of Artificial Neural Networks in Image Processing V. San Jose, California, [7] Darrel Hankerson, G.A. Harris and Peter D. Johnson, Jr. Introduction to information theory and data compression. CRC press, 1998 [8] [9] Chang W., H.S. Soliman et al. Preserving visual perception by learning clustering. ISCA: 1993 International Conference on Parallel and Distributed Computing and Systems, Louisville, Kent [10] Chang W., H.S. Soliman et al. A learning vector quantization neural model for image data compression. IEEE Data compression conference. 1994, Snowbird, Utah. [11] T. Kohonen, Kangas et al. LVQ_PAK: a program package for the correct application of learning vector quantization algorithms, IEEE International Joint Conference on Neural Networks [12] Erwin E., K. Obermayer et al. Self-organizing maps: ordering, convergence properties and energy functions. Biological Cybernetics 67: 1992 p35-45.

11 Divide the image into vector subimages, and reset the codebook occupied entries counter (COE=0). Set L to be the maximum number on entries in the codebook. Present an input vector subimage S Among the COE existing centroids, find the possible-winning-centroids set, PWCS, with an acceptable pixel differences with S, within the thresold Is PWCS empty? Yes COE L? Yes No Among the PWCS, and based on the Euclidian distance method, find j the index of the closest centroid to S. Consider the entire codebook to be PWCS. N Add S to the j s cluster. update its center of mass Cj to be the average of the new-formed cluster, only if the cluster is still of reasonable size (less than MaxTrainSet). Add new cluster j (COE=COE+1), with S as its centroid. Save j into the indices table, to be the representative index of S. Yes More subimages to train? No Construct the compressed file by appending the codebook entries to the end of the indices file, and zip it using a run length compression technique (LZW). Figure 1: DC training-compression flowchart.

12 Input vector subimages domain Sequentially presenting subimages Cluster Searching Searching Results Training Adding a Cluster (if need) Compression Saving (in order) the index of subimage s cluster Codebook Indices Figure 2: Training-compression phase. Load the compressed file, unzip it using LZW, and extract the sequence of indices and the codebook table. Concatenate, and compress using LZW Compressed File Point at the beginning of the indices sequence Get the next centroid index i Use i to look up centroid C i in the codebook table. Place, in order, the obtained C i in the decompressed image file, replacing the original corresponding subimage S i. Yes More indices to process? No End of the decompression process. Figure 3: DC decompression flowchart. Decompress using LZW, and deconcatenate Compressed File Indices Codebook Sequentially extracting indices Lookup an index Get the corresponding centroid Decompression Saving (in order) the extracted centroids Reconstructed Image Figure 4: Decompression phase.

13 L = 33 L = 256 L = 1024 MaxIntDif = 0 MaxIntDif = 50 MaxIntDif = 200 MaxTrainSet = 1 MaxTrainSet = 10 MaxTrainSet = 200 Figure 5: Quality changing based on different parameters values.

14 Gzip Comp. Ratio: :1 DjVu Comp. Ratio: 2,125.53:1 DC Comp. Ratio: 11,565.40:1 JPEG Comp. Ratio: 53.10:1 Quality: Foggy DjVu Comp. Ratio: 55.75:1 Quality: Foggy DC Comp. Ratio: :1 Quality: Lossless Figure 6: Background compressions using DC, JPEG, DjVu and Gzip.

15 Image DjVu DC Comp. Ratio Quality Comp. Ratio Quality Cdoc4.ppm Good Cdoc6.ppm Good Cdoc7.ppm Good Cdoc10.ppm Good Cdoc12.ppm Good Lossless Cdoc15.ppm Foggy Cdoc19.ppm Good Cdoc21.ppm Good Cdoc22.ppm Good Cdoc23.ppm Good Cdoc24.ppm Good Cdoc26.ppm Foggy Cdoc27.ppm Good Cdoc32.ppm Foggy Cdoc39.ppm Good Cdoc41.ppm Foggy Cdoc42.ppm Foggy Cdoc43.ppm Good Cdoc44.ppm Good Cdoc45.ppm Foggy Cdoc47.ppm Foggy Cdoc48.ppm Foggy Table 1: Compression results using DC and DjVu.

16 Cdoc10.ppm (Original) Compressed with DC Compressed with DjVu Cdoc12.ppm (Original) Compressed with DC Compressed with DjVu Cdoc29.ppm (Original) Compressed with DC Compressed with DjVu Figure 7: Examples of decompressed documents using DC and DjVu.

A Image Comparative Study using DCT, Fast Fourier, Wavelet Transforms and Huffman Algorithm

A Image Comparative Study using DCT, Fast Fourier, Wavelet Transforms and Huffman Algorithm International Journal of Engineering Research and General Science Volume 3, Issue 4, July-August, 15 ISSN 91-2730 A Image Comparative Study using DCT, Fast Fourier, Wavelet Transforms and Huffman Algorithm

More information

Optimization of Bit Rate in Medical Image Compression

Optimization of Bit Rate in Medical Image Compression Optimization of Bit Rate in Medical Image Compression Dr.J.Subash Chandra Bose 1, Mrs.Yamini.J 2, P.Pushparaj 3, P.Naveenkumar 4, Arunkumar.M 5, J.Vinothkumar 6 Professor and Head, Department of CSE, Professional

More information

Scalable Coding of Image Collections with Embedded Descriptors

Scalable Coding of Image Collections with Embedded Descriptors Scalable Coding of Image Collections with Embedded Descriptors N. Adami, A. Boschetti, R. Leonardi, P. Migliorati Department of Electronic for Automation, University of Brescia Via Branze, 38, Brescia,

More information

Image Compression Algorithm and JPEG Standard

Image Compression Algorithm and JPEG Standard International Journal of Scientific and Research Publications, Volume 7, Issue 12, December 2017 150 Image Compression Algorithm and JPEG Standard Suman Kunwar sumn2u@gmail.com Summary. The interest in

More information

Three Dimensional Motion Vectorless Compression

Three Dimensional Motion Vectorless Compression 384 IJCSNS International Journal of Computer Science and Network Security, VOL.9 No.4, April 9 Three Dimensional Motion Vectorless Compression Rohini Nagapadma and Narasimha Kaulgud* Department of E &

More information

Compression of Stereo Images using a Huffman-Zip Scheme

Compression of Stereo Images using a Huffman-Zip Scheme Compression of Stereo Images using a Huffman-Zip Scheme John Hamann, Vickey Yeh Department of Electrical Engineering, Stanford University Stanford, CA 94304 jhamann@stanford.edu, vickey@stanford.edu Abstract

More information

Interactive Progressive Encoding System For Transmission of Complex Images

Interactive Progressive Encoding System For Transmission of Complex Images Interactive Progressive Encoding System For Transmission of Complex Images Borko Furht 1, Yingli Wang 1, and Joe Celli 2 1 NSF Multimedia Laboratory Florida Atlantic University, Boca Raton, Florida 33431

More information

Data Compression. The Encoder and PCA

Data Compression. The Encoder and PCA Data Compression The Encoder and PCA Neural network techniques have been shown useful in the area of data compression. In general, data compression can be lossless compression or lossy compression. In

More information

CHAPTER 6 COUNTER PROPAGATION NEURAL NETWORK IN GAIT RECOGNITION

CHAPTER 6 COUNTER PROPAGATION NEURAL NETWORK IN GAIT RECOGNITION 75 CHAPTER 6 COUNTER PROPAGATION NEURAL NETWORK IN GAIT RECOGNITION 6.1 INTRODUCTION Counter propagation network (CPN) was developed by Robert Hecht-Nielsen as a means to combine an unsupervised Kohonen

More information

Figure (5) Kohonen Self-Organized Map

Figure (5) Kohonen Self-Organized Map 2- KOHONEN SELF-ORGANIZING MAPS (SOM) - The self-organizing neural networks assume a topological structure among the cluster units. - There are m cluster units, arranged in a one- or two-dimensional array;

More information

CHAPTER 6. 6 Huffman Coding Based Image Compression Using Complex Wavelet Transform. 6.3 Wavelet Transform based compression technique 106

CHAPTER 6. 6 Huffman Coding Based Image Compression Using Complex Wavelet Transform. 6.3 Wavelet Transform based compression technique 106 CHAPTER 6 6 Huffman Coding Based Image Compression Using Complex Wavelet Transform Page No 6.1 Introduction 103 6.2 Compression Techniques 104 103 6.2.1 Lossless compression 105 6.2.2 Lossy compression

More information

Digital Image Processing

Digital Image Processing Lecture 9+10 Image Compression Lecturer: Ha Dai Duong Faculty of Information Technology 1. Introduction Image compression To Solve the problem of reduncing the amount of data required to represent a digital

More information

PATTERN RECOGNITION USING NEURAL NETWORKS

PATTERN RECOGNITION USING NEURAL NETWORKS PATTERN RECOGNITION USING NEURAL NETWORKS Santaji Ghorpade 1, Jayshree Ghorpade 2 and Shamla Mantri 3 1 Department of Information Technology Engineering, Pune University, India santaji_11jan@yahoo.co.in,

More information

Dictionary Based Compression for Images

Dictionary Based Compression for Images Dictionary Based Compression for Images Bruno Carpentieri Abstract Lempel-Ziv methods were original introduced to compress one-dimensional data (text, object codes, etc.) but recently they have been successfully

More information

Image Compression for Mobile Devices using Prediction and Direct Coding Approach

Image Compression for Mobile Devices using Prediction and Direct Coding Approach Image Compression for Mobile Devices using Prediction and Direct Coding Approach Joshua Rajah Devadason M.E. scholar, CIT Coimbatore, India Mr. T. Ramraj Assistant Professor, CIT Coimbatore, India Abstract

More information

Chapter 7: Competitive learning, clustering, and self-organizing maps

Chapter 7: Competitive learning, clustering, and self-organizing maps Chapter 7: Competitive learning, clustering, and self-organizing maps António R. C. Paiva EEL 6814 Spring 2008 Outline Competitive learning Clustering Self-Organizing Maps What is competition in neural

More information

Modified SPIHT Image Coder For Wireless Communication

Modified SPIHT Image Coder For Wireless Communication Modified SPIHT Image Coder For Wireless Communication M. B. I. REAZ, M. AKTER, F. MOHD-YASIN Faculty of Engineering Multimedia University 63100 Cyberjaya, Selangor Malaysia Abstract: - The Set Partitioning

More information

REVIEW ON IMAGE COMPRESSION TECHNIQUES AND ADVANTAGES OF IMAGE COMPRESSION

REVIEW ON IMAGE COMPRESSION TECHNIQUES AND ADVANTAGES OF IMAGE COMPRESSION REVIEW ON IMAGE COMPRESSION TECHNIQUES AND ABSTRACT ADVANTAGES OF IMAGE COMPRESSION Amanpreet Kaur 1, Dr. Jagroop Singh 2 1 Ph. D Scholar, Deptt. of Computer Applications, IK Gujral Punjab Technical University,

More information

A new predictive image compression scheme using histogram analysis and pattern matching

A new predictive image compression scheme using histogram analysis and pattern matching University of Wollongong Research Online University of Wollongong in Dubai - Papers University of Wollongong in Dubai 00 A new predictive image compression scheme using histogram analysis and pattern matching

More information

Available online at ScienceDirect. Procedia Computer Science 89 (2016 )

Available online at   ScienceDirect. Procedia Computer Science 89 (2016 ) Available online at www.sciencedirect.com ScienceDirect Procedia Computer Science 89 (2016 ) 778 784 Twelfth International Multi-Conference on Information Processing-2016 (IMCIP-2016) Color Image Compression

More information

Fractal Compression. Related Topic Report. Henry Xiao. Queen s University. Kingston, Ontario, Canada. April 2004

Fractal Compression. Related Topic Report. Henry Xiao. Queen s University. Kingston, Ontario, Canada. April 2004 Fractal Compression Related Topic Report By Henry Xiao Queen s University Kingston, Ontario, Canada April 2004 Fractal Introduction Fractal is first introduced in geometry field. The birth of fractal geometry

More information

IMAGE PROCESSING (RRY025) LECTURE 13 IMAGE COMPRESSION - I

IMAGE PROCESSING (RRY025) LECTURE 13 IMAGE COMPRESSION - I IMAGE PROCESSING (RRY025) LECTURE 13 IMAGE COMPRESSION - I 1 Need For Compression 2D data sets are much larger than 1D. TV and movie data sets are effectively 3D (2-space, 1-time). Need Compression for

More information

A Novel Image Compression Technique using Simple Arithmetic Addition

A Novel Image Compression Technique using Simple Arithmetic Addition Proc. of Int. Conf. on Recent Trends in Information, Telecommunication and Computing, ITC A Novel Image Compression Technique using Simple Arithmetic Addition Nadeem Akhtar, Gufran Siddiqui and Salman

More information

ROI Based Image Compression in Baseline JPEG

ROI Based Image Compression in Baseline JPEG 168-173 RESEARCH ARTICLE OPEN ACCESS ROI Based Image Compression in Baseline JPEG M M M Kumar Varma #1, Madhuri. Bagadi #2 Associate professor 1, M.Tech Student 2 Sri Sivani College of Engineering, Department

More information

Research Article Does an Arithmetic Coding Followed by Run-length Coding Enhance the Compression Ratio?

Research Article Does an Arithmetic Coding Followed by Run-length Coding Enhance the Compression Ratio? Research Journal of Applied Sciences, Engineering and Technology 10(7): 736-741, 2015 DOI:10.19026/rjaset.10.2425 ISSN: 2040-7459; e-issn: 2040-7467 2015 Maxwell Scientific Publication Corp. Submitted:

More information

Review and Implementation of DWT based Scalable Video Coding with Scalable Motion Coding.

Review and Implementation of DWT based Scalable Video Coding with Scalable Motion Coding. Project Title: Review and Implementation of DWT based Scalable Video Coding with Scalable Motion Coding. Midterm Report CS 584 Multimedia Communications Submitted by: Syed Jawwad Bukhari 2004-03-0028 About

More information

SVM-based Filter Using Evidence Theory and Neural Network for Image Denosing

SVM-based Filter Using Evidence Theory and Neural Network for Image Denosing Journal of Software Engineering and Applications 013 6 106-110 doi:10.436/sea.013.63b03 Published Online March 013 (http://www.scirp.org/ournal/sea) SVM-based Filter Using Evidence Theory and Neural Network

More information

Function approximation using RBF network. 10 basis functions and 25 data points.

Function approximation using RBF network. 10 basis functions and 25 data points. 1 Function approximation using RBF network F (x j ) = m 1 w i ϕ( x j t i ) i=1 j = 1... N, m 1 = 10, N = 25 10 basis functions and 25 data points. Basis function centers are plotted with circles and data

More information

MRT based Fixed Block size Transform Coding

MRT based Fixed Block size Transform Coding 3 MRT based Fixed Block size Transform Coding Contents 3.1 Transform Coding..64 3.1.1 Transform Selection...65 3.1.2 Sub-image size selection... 66 3.1.3 Bit Allocation.....67 3.2 Transform coding using

More information

Part 1 of 4. MARCH

Part 1 of 4. MARCH Presented by Brought to You by Part 1 of 4 MARCH 2004 www.securitysales.com A1 Part1of 4 Essentials of DIGITAL VIDEO COMPRESSION By Bob Wimmer Video Security Consultants cctvbob@aol.com AT A GLANCE Compression

More information

AN ANALYTICAL STUDY OF LOSSY COMPRESSION TECHINIQUES ON CONTINUOUS TONE GRAPHICAL IMAGES

AN ANALYTICAL STUDY OF LOSSY COMPRESSION TECHINIQUES ON CONTINUOUS TONE GRAPHICAL IMAGES AN ANALYTICAL STUDY OF LOSSY COMPRESSION TECHINIQUES ON CONTINUOUS TONE GRAPHICAL IMAGES Dr.S.Narayanan Computer Centre, Alagappa University, Karaikudi-South (India) ABSTRACT The programs using complex

More information

IMAGE COMPRESSION. Chapter - 5 : (Basic)

IMAGE COMPRESSION. Chapter - 5 : (Basic) Chapter - 5 : IMAGE COMPRESSION (Basic) Q() Explain the different types of redundncies that exists in image.? (8M May6 Comp) [8M, MAY 7, ETRX] A common characteristic of most images is that the neighboring

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

Joint Image Classification and Compression Using Hierarchical Table-Lookup Vector Quantization

Joint Image Classification and Compression Using Hierarchical Table-Lookup Vector Quantization Joint Image Classification and Compression Using Hierarchical Table-Lookup Vector Quantization Navin Chadda, Keren Perlmuter and Robert M. Gray Information Systems Laboratory Stanford University CA-934305

More information

A COMPRESSION TECHNIQUES IN DIGITAL IMAGE PROCESSING - REVIEW

A COMPRESSION TECHNIQUES IN DIGITAL IMAGE PROCESSING - REVIEW A COMPRESSION TECHNIQUES IN DIGITAL IMAGE PROCESSING - ABSTRACT: REVIEW M.JEYAPRATHA 1, B.POORNA VENNILA 2 Department of Computer Application, Nadar Saraswathi College of Arts and Science, Theni, Tamil

More information

IMAGE COMPRESSION TECHNIQUES

IMAGE COMPRESSION TECHNIQUES IMAGE COMPRESSION TECHNIQUES A.VASANTHAKUMARI, M.Sc., M.Phil., ASSISTANT PROFESSOR OF COMPUTER SCIENCE, JOSEPH ARTS AND SCIENCE COLLEGE, TIRUNAVALUR, VILLUPURAM (DT), TAMIL NADU, INDIA ABSTRACT A picture

More information

Seismic regionalization based on an artificial neural network

Seismic regionalization based on an artificial neural network Seismic regionalization based on an artificial neural network *Jaime García-Pérez 1) and René Riaño 2) 1), 2) Instituto de Ingeniería, UNAM, CU, Coyoacán, México D.F., 014510, Mexico 1) jgap@pumas.ii.unam.mx

More information

CS 335 Graphics and Multimedia. Image Compression

CS 335 Graphics and Multimedia. Image Compression CS 335 Graphics and Multimedia Image Compression CCITT Image Storage and Compression Group 3: Huffman-type encoding for binary (bilevel) data: FAX Group 4: Entropy encoding without error checks of group

More information

Image Compression: An Artificial Neural Network Approach

Image Compression: An Artificial Neural Network Approach Image Compression: An Artificial Neural Network Approach Anjana B 1, Mrs Shreeja R 2 1 Department of Computer Science and Engineering, Calicut University, Kuttippuram 2 Department of Computer Science and

More information

Video Compression An Introduction

Video Compression An Introduction Video Compression An Introduction The increasing demand to incorporate video data into telecommunications services, the corporate environment, the entertainment industry, and even at home has made digital

More information

MRT based Adaptive Transform Coder with Classified Vector Quantization (MATC-CVQ)

MRT based Adaptive Transform Coder with Classified Vector Quantization (MATC-CVQ) 5 MRT based Adaptive Transform Coder with Classified Vector Quantization (MATC-CVQ) Contents 5.1 Introduction.128 5.2 Vector Quantization in MRT Domain Using Isometric Transformations and Scaling.130 5.2.1

More information

Face Recognition Using Vector Quantization Histogram and Support Vector Machine Classifier Rong-sheng LI, Fei-fei LEE *, Yan YAN and Qiu CHEN

Face Recognition Using Vector Quantization Histogram and Support Vector Machine Classifier Rong-sheng LI, Fei-fei LEE *, Yan YAN and Qiu CHEN 2016 International Conference on Artificial Intelligence: Techniques and Applications (AITA 2016) ISBN: 978-1-60595-389-2 Face Recognition Using Vector Quantization Histogram and Support Vector Machine

More information

CHAPTER 3 TUMOR DETECTION BASED ON NEURO-FUZZY TECHNIQUE

CHAPTER 3 TUMOR DETECTION BASED ON NEURO-FUZZY TECHNIQUE 32 CHAPTER 3 TUMOR DETECTION BASED ON NEURO-FUZZY TECHNIQUE 3.1 INTRODUCTION In this chapter we present the real time implementation of an artificial neural network based on fuzzy segmentation process

More information

Multimedia Communications. Transform Coding

Multimedia Communications. Transform Coding Multimedia Communications Transform Coding Transform coding Transform coding: source output is transformed into components that are coded according to their characteristics If a sequence of inputs is transformed

More information

CODING METHOD FOR EMBEDDING AUDIO IN VIDEO STREAM. Harri Sorokin, Jari Koivusaari, Moncef Gabbouj, and Jarmo Takala

CODING METHOD FOR EMBEDDING AUDIO IN VIDEO STREAM. Harri Sorokin, Jari Koivusaari, Moncef Gabbouj, and Jarmo Takala CODING METHOD FOR EMBEDDING AUDIO IN VIDEO STREAM Harri Sorokin, Jari Koivusaari, Moncef Gabbouj, and Jarmo Takala Tampere University of Technology Korkeakoulunkatu 1, 720 Tampere, Finland ABSTRACT In

More information

Index. 1. Motivation 2. Background 3. JPEG Compression The Discrete Cosine Transformation Quantization Coding 4. MPEG 5.

Index. 1. Motivation 2. Background 3. JPEG Compression The Discrete Cosine Transformation Quantization Coding 4. MPEG 5. Index 1. Motivation 2. Background 3. JPEG Compression The Discrete Cosine Transformation Quantization Coding 4. MPEG 5. Literature Lossy Compression Motivation To meet a given target bit-rate for storage

More information

Differential Compression and Optimal Caching Methods for Content-Based Image Search Systems

Differential Compression and Optimal Caching Methods for Content-Based Image Search Systems Differential Compression and Optimal Caching Methods for Content-Based Image Search Systems Di Zhong a, Shih-Fu Chang a, John R. Smith b a Department of Electrical Engineering, Columbia University, NY,

More information

A Comparative Study of Conventional and Neural Network Classification of Multispectral Data

A Comparative Study of Conventional and Neural Network Classification of Multispectral Data A Comparative Study of Conventional and Neural Network Classification of Multispectral Data B.Solaiman & M.C.Mouchot Ecole Nationale Supérieure des Télécommunications de Bretagne B.P. 832, 29285 BREST

More information

Extract an Essential Skeleton of a Character as a Graph from a Character Image

Extract an Essential Skeleton of a Character as a Graph from a Character Image Extract an Essential Skeleton of a Character as a Graph from a Character Image Kazuhisa Fujita University of Electro-Communications 1-5-1 Chofugaoka, Chofu, Tokyo, 182-8585 Japan k-z@nerve.pc.uec.ac.jp

More information

IMAGE COMPRESSION. Image Compression. Why? Reducing transportation times Reducing file size. A two way event - compression and decompression

IMAGE COMPRESSION. Image Compression. Why? Reducing transportation times Reducing file size. A two way event - compression and decompression IMAGE COMPRESSION Image Compression Why? Reducing transportation times Reducing file size A two way event - compression and decompression 1 Compression categories Compression = Image coding Still-image

More information

Statistical Image Compression using Fast Fourier Coefficients

Statistical Image Compression using Fast Fourier Coefficients Statistical Image Compression using Fast Fourier Coefficients M. Kanaka Reddy Research Scholar Dept.of Statistics Osmania University Hyderabad-500007 V. V. Haragopal Professor Dept.of Statistics Osmania

More information

Features. Sequential encoding. Progressive encoding. Hierarchical encoding. Lossless encoding using a different strategy

Features. Sequential encoding. Progressive encoding. Hierarchical encoding. Lossless encoding using a different strategy JPEG JPEG Joint Photographic Expert Group Voted as international standard in 1992 Works with color and grayscale images, e.g., satellite, medical,... Motivation: The compression ratio of lossless methods

More information

7.5 Dictionary-based Coding

7.5 Dictionary-based Coding 7.5 Dictionary-based Coding LZW uses fixed-length code words to represent variable-length strings of symbols/characters that commonly occur together, e.g., words in English text LZW encoder and decoder

More information

DIGITAL TELEVISION 1. DIGITAL VIDEO FUNDAMENTALS

DIGITAL TELEVISION 1. DIGITAL VIDEO FUNDAMENTALS DIGITAL TELEVISION 1. DIGITAL VIDEO FUNDAMENTALS Television services in Europe currently broadcast video at a frame rate of 25 Hz. Each frame consists of two interlaced fields, giving a field rate of 50

More information

1. Introduction. 2. Motivation and Problem Definition. Volume 8 Issue 2, February Susmita Mohapatra

1. Introduction. 2. Motivation and Problem Definition. Volume 8 Issue 2, February Susmita Mohapatra Pattern Recall Analysis of the Hopfield Neural Network with a Genetic Algorithm Susmita Mohapatra Department of Computer Science, Utkal University, India Abstract: This paper is focused on the implementation

More information

VIDEO SIGNALS. Lossless coding

VIDEO SIGNALS. Lossless coding VIDEO SIGNALS Lossless coding LOSSLESS CODING The goal of lossless image compression is to represent an image signal with the smallest possible number of bits without loss of any information, thereby speeding

More information

Digital Image Representation Image Compression

Digital Image Representation Image Compression Digital Image Representation Image Compression 1 Image Representation Standards Need for compression Compression types Lossless compression Lossy compression Image Compression Basics Redundancy/redundancy

More information

Adaptive Quantization for Video Compression in Frequency Domain

Adaptive Quantization for Video Compression in Frequency Domain Adaptive Quantization for Video Compression in Frequency Domain *Aree A. Mohammed and **Alan A. Abdulla * Computer Science Department ** Mathematic Department University of Sulaimani P.O.Box: 334 Sulaimani

More information

Image Compression with Competitive Networks and Pre-fixed Prototypes*

Image Compression with Competitive Networks and Pre-fixed Prototypes* Image Compression with Competitive Networks and Pre-fixed Prototypes* Enrique Merida-Casermeiro^, Domingo Lopez-Rodriguez^, and Juan M. Ortiz-de-Lazcano-Lobato^ ^ Department of Applied Mathematics, University

More information

Lecture 5: Compression I. This Week s Schedule

Lecture 5: Compression I. This Week s Schedule Lecture 5: Compression I Reading: book chapter 6, section 3 &5 chapter 7, section 1, 2, 3, 4, 8 Today: This Week s Schedule The concept behind compression Rate distortion theory Image compression via DCT

More information

IMAGE COMPRESSION USING HYBRID QUANTIZATION METHOD IN JPEG

IMAGE COMPRESSION USING HYBRID QUANTIZATION METHOD IN JPEG IMAGE COMPRESSION USING HYBRID QUANTIZATION METHOD IN JPEG MANGESH JADHAV a, SNEHA GHANEKAR b, JIGAR JAIN c a 13/A Krishi Housing Society, Gokhale Nagar, Pune 411016,Maharashtra, India. (mail2mangeshjadhav@gmail.com)

More information

A Reversible Data Hiding Scheme for BTC- Compressed Images

A Reversible Data Hiding Scheme for BTC- Compressed Images IJACSA International Journal of Advanced Computer Science and Applications, A Reversible Data Hiding Scheme for BTC- Compressed Images Ching-Chiuan Lin Shih-Chieh Chen Department of Multimedia and Game

More information

THREE DESCRIPTIONS OF SCALAR QUANTIZATION SYSTEM FOR EFFICIENT DATA TRANSMISSION

THREE DESCRIPTIONS OF SCALAR QUANTIZATION SYSTEM FOR EFFICIENT DATA TRANSMISSION THREE DESCRIPTIONS OF SCALAR QUANTIZATION SYSTEM FOR EFFICIENT DATA TRANSMISSION Hui Ting Teo and Mohd Fadzli bin Mohd Salleh School of Electrical and Electronic Engineering Universiti Sains Malaysia,

More information

A Very Low Bit Rate Image Compressor Using Transformed Classified Vector Quantization

A Very Low Bit Rate Image Compressor Using Transformed Classified Vector Quantization Informatica 29 (2005) 335 341 335 A Very Low Bit Rate Image Compressor Using Transformed Classified Vector Quantization Hsien-Wen Tseng Department of Information Management Chaoyang University of Technology

More information

Vector Quantization of Images Using Modified Adaptive Resonance Algorithm for Hierarchical Clustering

Vector Quantization of Images Using Modified Adaptive Resonance Algorithm for Hierarchical Clustering IEEE TRANSACTIONS ON NEURAL NETWORKS, VOL. 12, NO. 5, SEPTEMBER 2001 1147 Vector Quantization of Images Using Modified Adaptive Resonance Algorithm for Hierarchical Clustering Natalija Vlajic and Howard

More information

Digital Image Steganography Techniques: Case Study. Karnataka, India.

Digital Image Steganography Techniques: Case Study. Karnataka, India. ISSN: 2320 8791 (Impact Factor: 1.479) Digital Image Steganography Techniques: Case Study Santosh Kumar.S 1, Archana.M 2 1 Department of Electronicsand Communication Engineering, Sri Venkateshwara College

More information

Image Compression using Single Layer Linear Neural Networks B.Arunapriya a, * D.Kavitha Devi b

Image Compression using Single Layer Linear Neural Networks B.Arunapriya a, * D.Kavitha Devi b Available online at www.sciencedirect.com Procedia Computer Science 00 (2009) 000 000 Procedia Computer Science 2 (2010) 345 352 ICEBT 2010 Procedia Computer Science www.elsevier.com/locate/procedia Image

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

A Comprehensive Review of Data Compression Techniques

A Comprehensive Review of Data Compression Techniques Volume-6, Issue-2, March-April 2016 International Journal of Engineering and Management Research Page Number: 684-688 A Comprehensive Review of Data Compression Techniques Palwinder Singh 1, Amarbir Singh

More information

HYBRID TRANSFORMATION TECHNIQUE FOR IMAGE COMPRESSION

HYBRID TRANSFORMATION TECHNIQUE FOR IMAGE COMPRESSION 31 st July 01. Vol. 41 No. 005-01 JATIT & LLS. All rights reserved. ISSN: 199-8645 www.jatit.org E-ISSN: 1817-3195 HYBRID TRANSFORMATION TECHNIQUE FOR IMAGE COMPRESSION 1 SRIRAM.B, THIYAGARAJAN.S 1, Student,

More information

So, what is data compression, and why do we need it?

So, what is data compression, and why do we need it? In the last decade we have been witnessing a revolution in the way we communicate 2 The major contributors in this revolution are: Internet; The explosive development of mobile communications; and The

More information

Fundamentals of Video Compression. Video Compression

Fundamentals of Video Compression. Video Compression Fundamentals of Video Compression Introduction to Digital Video Basic Compression Techniques Still Image Compression Techniques - JPEG Video Compression Introduction to Digital Video Video is a stream

More information

ECE 499/599 Data Compression & Information Theory. Thinh Nguyen Oregon State University

ECE 499/599 Data Compression & Information Theory. Thinh Nguyen Oregon State University ECE 499/599 Data Compression & Information Theory Thinh Nguyen Oregon State University Adminstrivia Office Hours TTh: 2-3 PM Kelley Engineering Center 3115 Class homepage http://www.eecs.orst.edu/~thinhq/teaching/ece499/spring06/spring06.html

More information

Fingerprint Image Compression

Fingerprint Image Compression Fingerprint Image Compression Ms.Mansi Kambli 1*,Ms.Shalini Bhatia 2 * Student 1*, Professor 2 * Thadomal Shahani Engineering College * 1,2 Abstract Modified Set Partitioning in Hierarchical Tree with

More information

A Review on Digital Image Compression Techniques

A Review on Digital Image Compression Techniques A Review on Digital Image Compression Techniques Er. Shilpa Sachdeva Yadwindra College of Engineering Talwandi Sabo,Punjab,India +91-9915719583 s.sachdeva88@gmail.com Er. Rajbhupinder Kaur Department of

More information

A LOSSLESS INDEX CODING ALGORITHM AND VLSI DESIGN FOR VECTOR QUANTIZATION

A LOSSLESS INDEX CODING ALGORITHM AND VLSI DESIGN FOR VECTOR QUANTIZATION A LOSSLESS INDEX CODING ALGORITHM AND VLSI DESIGN FOR VECTOR QUANTIZATION Ming-Hwa Sheu, Sh-Chi Tsai and Ming-Der Shieh Dept. of Electronic Eng., National Yunlin Univ. of Science and Technology, Yunlin,

More information

Hybrid Image Compression Using DWT, DCT and Huffman Coding. Techniques

Hybrid Image Compression Using DWT, DCT and Huffman Coding. Techniques Hybrid Image Compression Using DWT, DCT and Huffman Coding Techniques Veerpal kaur, Gurwinder kaur Abstract- Here in this hybrid model we are going to proposed a Nobel technique which is the combination

More information

IMAGE CLASSIFICATION USING COMPETITIVE NEURAL NETWORKS

IMAGE CLASSIFICATION USING COMPETITIVE NEURAL NETWORKS IMAGE CLASSIFICATION USING COMPETITIVE NEURAL NETWORKS V. Musoko, M. Kolı nova, A. Procha zka Institute of Chemical Technology, Department of Computing and Control Engineering Abstract The contribution

More information

Final Review. Image Processing CSE 166 Lecture 18

Final Review. Image Processing CSE 166 Lecture 18 Final Review Image Processing CSE 166 Lecture 18 Topics covered Basis vectors Matrix based transforms Wavelet transform Image compression Image watermarking Morphological image processing Segmentation

More information

Topic 5 Image Compression

Topic 5 Image Compression Topic 5 Image Compression Introduction Data Compression: The process of reducing the amount of data required to represent a given quantity of information. Purpose of Image Compression: the reduction of

More information

Information Fusion Dr. B. K. Panigrahi

Information Fusion Dr. B. K. Panigrahi Information Fusion By Dr. B. K. Panigrahi Asst. Professor Department of Electrical Engineering IIT Delhi, New Delhi-110016 01/12/2007 1 Introduction Classification OUTLINE K-fold cross Validation Feature

More information

Module 8: Video Coding Basics Lecture 42: Sub-band coding, Second generation coding, 3D coding. The Lecture Contains: Performance Measures

Module 8: Video Coding Basics Lecture 42: Sub-band coding, Second generation coding, 3D coding. The Lecture Contains: Performance Measures The Lecture Contains: Performance Measures file:///d /...Ganesh%20Rana)/MY%20COURSE_Ganesh%20Rana/Prof.%20Sumana%20Gupta/FINAL%20DVSP/lecture%2042/42_1.htm[12/31/2015 11:57:52 AM] 3) Subband Coding It

More information

KeyWords: Image Compression, LBG, ENN, BPNN, FBP.

KeyWords: Image Compression, LBG, ENN, BPNN, FBP. Volume 3, Issue 6, June 2013 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com Image Compression

More information

FRACTAL IMAGE COMPRESSION OF GRAYSCALE AND RGB IMAGES USING DCT WITH QUADTREE DECOMPOSITION AND HUFFMAN CODING. Moheb R. Girgis and Mohammed M.

FRACTAL IMAGE COMPRESSION OF GRAYSCALE AND RGB IMAGES USING DCT WITH QUADTREE DECOMPOSITION AND HUFFMAN CODING. Moheb R. Girgis and Mohammed M. 322 FRACTAL IMAGE COMPRESSION OF GRAYSCALE AND RGB IMAGES USING DCT WITH QUADTREE DECOMPOSITION AND HUFFMAN CODING Moheb R. Girgis and Mohammed M. Talaat Abstract: Fractal image compression (FIC) is a

More information

New wavelet based ART network for texture classification

New wavelet based ART network for texture classification University of Wollongong Research Online Faculty of Informatics - Papers (Archive) Faculty of Engineering and Information Sciences 1996 New wavelet based ART network for texture classification Jiazhao

More information

Pattern Recognition. Kjell Elenius. Speech, Music and Hearing KTH. March 29, 2007 Speech recognition

Pattern Recognition. Kjell Elenius. Speech, Music and Hearing KTH. March 29, 2007 Speech recognition Pattern Recognition Kjell Elenius Speech, Music and Hearing KTH March 29, 2007 Speech recognition 2007 1 Ch 4. Pattern Recognition 1(3) Bayes Decision Theory Minimum-Error-Rate Decision Rules Discriminant

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

Large-scale visual recognition Efficient matching

Large-scale visual recognition Efficient matching Large-scale visual recognition Efficient matching Florent Perronnin, XRCE Hervé Jégou, INRIA CVPR tutorial June 16, 2012 Outline!! Preliminary!! Locality Sensitive Hashing: the two modes!! Hashing!! Embedding!!

More information

Image Compression With Haar Discrete Wavelet Transform

Image Compression With Haar Discrete Wavelet Transform Image Compression With Haar Discrete Wavelet Transform Cory Cox ME 535: Computational Techniques in Mech. Eng. Figure 1 : An example of the 2D discrete wavelet transform that is used in JPEG2000. Source:

More information

Slide07 Haykin Chapter 9: Self-Organizing Maps

Slide07 Haykin Chapter 9: Self-Organizing Maps Slide07 Haykin Chapter 9: Self-Organizing Maps CPSC 636-600 Instructor: Yoonsuck Choe Spring 2012 Introduction Self-organizing maps (SOM) is based on competitive learning, where output neurons compete

More information

Wireless Communication

Wireless Communication Wireless Communication Systems @CS.NCTU Lecture 6: Image Instructor: Kate Ching-Ju Lin ( 林靖茹 ) Chap. 9 of Fundamentals of Multimedia Some reference from http://media.ee.ntu.edu.tw/courses/dvt/15f/ 1 Outline

More information

Use of Local Minimization for Lossless Gray Image Compression

Use of Local Minimization for Lossless Gray Image Compression Narendra Kumar et al. / (IJCSIT) International Journal of Computer Science and Information Technologies, Vol. 1 Use of Local Minimization for Lossless Gray Image Compression Narendra Kumar 1, Dr. Sachin

More information

A Generalized Method to Solve Text-Based CAPTCHAs

A Generalized Method to Solve Text-Based CAPTCHAs A Generalized Method to Solve Text-Based CAPTCHAs Jason Ma, Bilal Badaoui, Emile Chamoun December 11, 2009 1 Abstract We present work in progress on the automated solving of text-based CAPTCHAs. Our method

More information

1.Define image compression. Explain about the redundancies in a digital image.

1.Define image compression. Explain about the redundancies in a digital image. 1.Define image compression. Explain about the redundancies in a digital image. The term data compression refers to the process of reducing the amount of data required to represent a given quantity of information.

More information

Neural Network based textural labeling of images in multimedia applications

Neural Network based textural labeling of images in multimedia applications Neural Network based textural labeling of images in multimedia applications S.A. Karkanis +, G.D. Magoulas +, and D.A. Karras ++ + University of Athens, Dept. of Informatics, Typa Build., Panepistimiopolis,

More information

IMPLEMENTATION OF RBF TYPE NETWORKS BY SIGMOIDAL FEEDFORWARD NEURAL NETWORKS

IMPLEMENTATION OF RBF TYPE NETWORKS BY SIGMOIDAL FEEDFORWARD NEURAL NETWORKS IMPLEMENTATION OF RBF TYPE NETWORKS BY SIGMOIDAL FEEDFORWARD NEURAL NETWORKS BOGDAN M.WILAMOWSKI University of Wyoming RICHARD C. JAEGER Auburn University ABSTRACT: It is shown that by introducing special

More information

CHAPTER 4 REVERSIBLE IMAGE WATERMARKING USING BIT PLANE CODING AND LIFTING WAVELET TRANSFORM

CHAPTER 4 REVERSIBLE IMAGE WATERMARKING USING BIT PLANE CODING AND LIFTING WAVELET TRANSFORM 74 CHAPTER 4 REVERSIBLE IMAGE WATERMARKING USING BIT PLANE CODING AND LIFTING WAVELET TRANSFORM Many data embedding methods use procedures that in which the original image is distorted by quite a small

More information

International Journal of Emerging Technology and Advanced Engineering Website: (ISSN , Volume 2, Issue 4, April 2012)

International Journal of Emerging Technology and Advanced Engineering Website:   (ISSN , Volume 2, Issue 4, April 2012) A Technical Analysis Towards Digital Video Compression Rutika Joshi 1, Rajesh Rai 2, Rajesh Nema 3 1 Student, Electronics and Communication Department, NIIST College, Bhopal, 2,3 Prof., Electronics and

More information

Data Compression. Media Signal Processing, Presentation 2. Presented By: Jahanzeb Farooq Michael Osadebey

Data Compression. Media Signal Processing, Presentation 2. Presented By: Jahanzeb Farooq Michael Osadebey Data Compression Media Signal Processing, Presentation 2 Presented By: Jahanzeb Farooq Michael Osadebey What is Data Compression? Definition -Reducing the amount of data required to represent a source

More information

Multimedia Systems Image III (Image Compression, JPEG) Mahdi Amiri April 2011 Sharif University of Technology

Multimedia Systems Image III (Image Compression, JPEG) Mahdi Amiri April 2011 Sharif University of Technology Course Presentation Multimedia Systems Image III (Image Compression, JPEG) Mahdi Amiri April 2011 Sharif University of Technology Image Compression Basics Large amount of data in digital images File size

More information