We utilize both the implicit residual source correlation and the explicit redundancy from a forward error correction (FEC) scheme for the error protection of packetized variable-length encoded source indices. The implicit source correlation is exploited in a novel symbol-based soft-input a-posteriori probability (APP) decoder, which leads to an optimal decoding process in combination with a mean-squares or maximum a-posteriori probability estimation of the reconstructed source signal. When, additionally, the variable-length encoded source data is protected by channel codes, an iterative source-channel decoder can be obtained in the same way as for serially concatenated codes, where the outer constituent decoder is replaced by the proposed APP source decoder. Simulation results show that, by additionally considering the correlations between the variable-length encoded source indices, the error-correction performance can be highly increased.
{"title":"Combining FEC and optimal soft-input source decoding for the reliable transmission of correlated variable-length encoded signals","authors":"J. Kliewer, R. Thobaben","doi":"10.1109/DCC.2002.999946","DOIUrl":"https://doi.org/10.1109/DCC.2002.999946","url":null,"abstract":"We utilize both the implicit residual source correlation and the explicit redundancy from a forward error correction (FEC) scheme for the error protection of packetized variable-length encoded source indices. The implicit source correlation is exploited in a novel symbol-based soft-input a-posteriori probability (APP) decoder, which leads to an optimal decoding process in combination with a mean-squares or maximum a-posteriori probability estimation of the reconstructed source signal. When, additionally, the variable-length encoded source data is protected by channel codes, an iterative source-channel decoder can be obtained in the same way as for serially concatenated codes, where the outer constituent decoder is replaced by the proposed APP source decoder. Simulation results show that, by additionally considering the correlations between the variable-length encoded source indices, the error-correction performance can be highly increased.","PeriodicalId":420897,"journal":{"name":"Proceedings DCC 2002. Data Compression Conference","volume":"81 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124881304","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A 2D version of PPM (Prediction by Partial Matching) coding is introduced simply by combining a 2D template with the standard PPM coding scheme. A simple scheme for resolution reduction is given and the 2D PPM scheme extended to resolution progressive coding by placing pixels in a lower resolution image layer. The resolution is increased by a factor of 2 in each step. The 2D PPM coding is applied to palette images and street maps. The sequential results are comparable to PWC. The PPM results are a little better for the palette images with few colors (up to 4-5 bpp) and a little worse for the images with more colors. For street maps the 2D PPM is slightly better. The PPM based resolution progressive coding provides a better result than coding the resolution layers as individual images. Compared to GIF the resolution progressive 2D PPM's coding efficiency is significantly better. An example of combined content-layer/spatial progressive coding is also given.
{"title":"Progressive coding of palette images and digital maps","authors":"S. Forchhammer, J. M. Salinas","doi":"10.1109/DCC.2002.999974","DOIUrl":"https://doi.org/10.1109/DCC.2002.999974","url":null,"abstract":"A 2D version of PPM (Prediction by Partial Matching) coding is introduced simply by combining a 2D template with the standard PPM coding scheme. A simple scheme for resolution reduction is given and the 2D PPM scheme extended to resolution progressive coding by placing pixels in a lower resolution image layer. The resolution is increased by a factor of 2 in each step. The 2D PPM coding is applied to palette images and street maps. The sequential results are comparable to PWC. The PPM results are a little better for the palette images with few colors (up to 4-5 bpp) and a little worse for the images with more colors. For street maps the 2D PPM is slightly better. The PPM based resolution progressive coding provides a better result than coding the resolution layers as individual images. Compared to GIF the resolution progressive 2D PPM's coding efficiency is significantly better. An example of combined content-layer/spatial progressive coding is also given.","PeriodicalId":420897,"journal":{"name":"Proceedings DCC 2002. Data Compression Conference","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129361277","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2002-04-02DOI: 10.1109/DCC.2002.1000006
Gwenaelle Marquant
Summary form only given. The developments in video coding research deal with solutions to improve the picture quality while decreasing the bit rates. However, no major breakthrough in compression emerged and low bit rate high quality video compression is still an open issue. The compression scheme is generally decomposed into two stages: coding and decoding. In order to improve the compression efficiency, a complementary solution may consist in introducing a preprocessing stage before the encoding process or/and a post-processing step after decoding. For this purpose, instead of using the usual (Y, U, V) representation space to compress the video signal, where the video is encoded along different separate channels (luminance Y, chrominance U, chrominance V), we propose to choose other channels by means of a color preprocessing based upon perceptual and physics-based approaches. We compare an original H.26L encoder (ITU standard for video coding), i.e. without preprocessing, and the same H.26L encoder with a preprocessing stage to evaluate the extent to which the preprocessing stage increases the compression efficiency, in particular with perceptual solutions.
只提供摘要形式。视频编码研究的发展涉及在降低码率的同时提高图像质量的解决方案。然而，在压缩方面没有出现重大突破，低比特率高质量视频压缩仍然是一个悬而未决的问题。压缩方案一般分为两个阶段:编码和解码。为了提高压缩效率，一种互补的解决方案可以包括在编码处理之前引入预处理阶段或/和在解码之后引入后处理步骤。为此，我们不使用通常的(Y, U, V)表示空间来压缩视频信号，其中视频沿着不同的独立通道(亮度Y，色度U，色度V)进行编码，我们建议通过基于感知和基于物理的方法的颜色预处理来选择其他通道。我们比较了原始的H.26L编码器(国际电联视频编码标准)，即没有预处理，以及具有预处理阶段的相同H.26L编码器，以评估预处理阶段提高压缩效率的程度，特别是在感知解决方案中。
{"title":"Perceptual preprocessing techniques applied to video compression: some result elements and analysis","authors":"Gwenaelle Marquant","doi":"10.1109/DCC.2002.1000006","DOIUrl":"https://doi.org/10.1109/DCC.2002.1000006","url":null,"abstract":"Summary form only given. The developments in video coding research deal with solutions to improve the picture quality while decreasing the bit rates. However, no major breakthrough in compression emerged and low bit rate high quality video compression is still an open issue. The compression scheme is generally decomposed into two stages: coding and decoding. In order to improve the compression efficiency, a complementary solution may consist in introducing a preprocessing stage before the encoding process or/and a post-processing step after decoding. For this purpose, instead of using the usual (Y, U, V) representation space to compress the video signal, where the video is encoded along different separate channels (luminance Y, chrominance U, chrominance V), we propose to choose other channels by means of a color preprocessing based upon perceptual and physics-based approaches. We compare an original H.26L encoder (ITU standard for video coding), i.e. without preprocessing, and the same H.26L encoder with a preprocessing stage to evaluate the extent to which the preprocessing stage increases the compression efficiency, in particular with perceptual solutions.","PeriodicalId":420897,"journal":{"name":"Proceedings DCC 2002. Data Compression Conference","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124692460","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
An algorithm for supervised classification using vector quantization and entropy coding is presented. The classification rule is formed from a set of training data {(X/sub i/, Y/sub i/)}/sub i=1//sup n/, which are independent samples from a joint distribution P/sub XY/. Based on the principle of minimum description length (MDL), a statistical model that approximates the distribution P/sub XY/ ought to enable efficient coding of X and Y. On the other hand, we expect a system that encodes (X, Y) efficiently to provide ample information on the distribution P/sub XY/. This information can then be used to classify X, i.e., to predict the corresponding Y based on X. To encode both X and Y, a two-stage vector quantizer is applied to X and a Huffman code is formed for Y conditioned on each quantized value of X. The optimization of the encoder is equivalent to the design of a vector quantizer with an objective function reflecting the joint penalty of quantization error and misclassification rate. This vector quantizer provides an estimation of the conditional distribution of Y given X, which in turn yields an approximation to the Bayes classification rule. This algorithm, namely discriminant vector quantization (DVQ), is compared with learning vector quantization (LVQ) and CART/sup R/ on a number of data sets. DVQ outperforms the other two on several data sets. The relation between DVQ, density estimation, and regression is also discussed.
{"title":"A source coding approach to classification by vector quantization and the principle of minimum description length","authors":"Jia Li","doi":"10.1109/DCC.2002.999978","DOIUrl":"https://doi.org/10.1109/DCC.2002.999978","url":null,"abstract":"An algorithm for supervised classification using vector quantization and entropy coding is presented. The classification rule is formed from a set of training data {(X/sub i/, Y/sub i/)}/sub i=1//sup n/, which are independent samples from a joint distribution P/sub XY/. Based on the principle of minimum description length (MDL), a statistical model that approximates the distribution P/sub XY/ ought to enable efficient coding of X and Y. On the other hand, we expect a system that encodes (X, Y) efficiently to provide ample information on the distribution P/sub XY/. This information can then be used to classify X, i.e., to predict the corresponding Y based on X. To encode both X and Y, a two-stage vector quantizer is applied to X and a Huffman code is formed for Y conditioned on each quantized value of X. The optimization of the encoder is equivalent to the design of a vector quantizer with an objective function reflecting the joint penalty of quantization error and misclassification rate. This vector quantizer provides an estimation of the conditional distribution of Y given X, which in turn yields an approximation to the Bayes classification rule. This algorithm, namely discriminant vector quantization (DVQ), is compared with learning vector quantization (LVQ) and CART/sup R/ on a number of data sets. DVQ outperforms the other two on several data sets. The relation between DVQ, density estimation, and regression is also discussed.","PeriodicalId":420897,"journal":{"name":"Proceedings DCC 2002. Data Compression Conference","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130321983","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2002-04-02DOI: 10.1109/DCC.2002.1000017
M. Titchener
Summary form only given. Titchener (see Proc. DCC00, IEEE Society Press, p.353-62, 2000, and IEEE-ISIT, , MIT, Boston, August 1998) defined a computable grammar-based entropy measure (T-entropy) for finite strings, Ebeling, Steuer and Titchener (see Stochastics and Dynamics, vol.1, no.1, 2000) and Titchener and Ebeling (see Proc. DCC01, IEEE Society Press, p.520, 2001) demonstrated against the known results for the logistic map, to be a practical way to compute the Shannon information content for data files. A range of binary encodings of the logistic map dynamics have been prepared from a generating bi-partition and with selected normalised entropies, 0.1-1.0 bits/symbol, in steps of 0.1. This corpus of ten test files has been used to evaluate the 'absolute' performance of a series of popular compressors.
只提供摘要形式。Titchener(参见Proc. DCC00, IEEE Society Press, p.353-62, 2000, IEEE- isit，麻省理工学院，波士顿，1998年8月)为有限字符串定义了一个可计算的基于语法的熵测度(t -熵)，Ebeling, Steuer和Titchener(参见《随机与动力学》，vol.1, no. 1)。1, 2000)和Titchener和Ebeling(见Proc. DCC01, IEEE Society Press, p.520, 2001)针对逻辑图的已知结果证明，这是计算数据文件香农信息内容的实用方法。逻辑映射动力学的一系列二进制编码已经从生成的双分区和选择的归一化熵中准备好，0.1-1.0比特/符号，步骤为0.1。这个由十个测试文件组成的语料库已被用于评估一系列流行压缩器的“绝对”性能。
{"title":"Compressor performance, absolutely!","authors":"M. Titchener","doi":"10.1109/DCC.2002.1000017","DOIUrl":"https://doi.org/10.1109/DCC.2002.1000017","url":null,"abstract":"Summary form only given. Titchener (see Proc. DCC00, IEEE Society Press, p.353-62, 2000, and IEEE-ISIT, , MIT, Boston, August 1998) defined a computable grammar-based entropy measure (T-entropy) for finite strings, Ebeling, Steuer and Titchener (see Stochastics and Dynamics, vol.1, no.1, 2000) and Titchener and Ebeling (see Proc. DCC01, IEEE Society Press, p.520, 2001) demonstrated against the known results for the logistic map, to be a practical way to compute the Shannon information content for data files. A range of binary encodings of the logistic map dynamics have been prepared from a generating bi-partition and with selected normalised entropies, 0.1-1.0 bits/symbol, in steps of 0.1. This corpus of ten test files has been used to evaluate the 'absolute' performance of a series of popular compressors.","PeriodicalId":420897,"journal":{"name":"Proceedings DCC 2002. Data Compression Conference","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126881573","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Summary form only given. BCJR based source coding of image residuals is investigated. From a trellis representation of the residual, a joint source-channel coding system is formed. Then the BCJR algorithm is applied to find the MAP encoding. MAP and minimized squared error encoding are compared. The novelty of this work is the use of the BCJR algorithm and the MAP criterion in the source coding procedure. The source encoding system described preserves more features than an MSE based encoder. Also, blocking artifacts are reduced. Comparisons may be found in the full paper version (see http://www.it.lth.se/tomas/eriksson/spl I.bar/novak/spl I.bar/anderson/spl I.bar/dcc02.ps, 2001).
{"title":"Image coding with the MAP criterion","authors":"T. Eriksson, John B. Anderson, M. Novak","doi":"10.1109/DCC.2002.999996","DOIUrl":"https://doi.org/10.1109/DCC.2002.999996","url":null,"abstract":"Summary form only given. BCJR based source coding of image residuals is investigated. From a trellis representation of the residual, a joint source-channel coding system is formed. Then the BCJR algorithm is applied to find the MAP encoding. MAP and minimized squared error encoding are compared. The novelty of this work is the use of the BCJR algorithm and the MAP criterion in the source coding procedure. The source encoding system described preserves more features than an MSE based encoder. Also, blocking artifacts are reduced. Comparisons may be found in the full paper version (see http://www.it.lth.se/tomas/eriksson/spl I.bar/novak/spl I.bar/anderson/spl I.bar/dcc02.ps, 2001).","PeriodicalId":420897,"journal":{"name":"Proceedings DCC 2002. Data Compression Conference","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129116564","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Summary form only given. Forward error correction (FEC) based schemes are use widely to address the packet loss problem for Internet video. Given total available bandwidth, finding optimal bit allocation is very important in FEC-based video, because the FEC bit rate limits the rate available to compress video. We want to give proper protection to the source, but also prevent unwanted FEC rate expansion. The rate of packet headers is often ignored in allocating bit rate. We show that this packetization overhead has significant influence on system performance in many cases. Decreasing packet size increases the rate of packet headers, thus reducing the available rate for the source and its FEC codes. On the other hand, smaller packet size allows a larger number of packets, in which case it can be shown that the efficiency of FEC codes improves. We show that packet size should be optimized to balance the effect of packet headers and the efficiency of FEC codes. We develop a probabilistic framework for the solution of rate allocation problem in the presence of packet overhead. We implement our solution on the MPEG-4 fine granularity scalability (FGS) mode. To show the flexibility of our technique, we use an unequal error protection scheme with FGS. Experimental results show that our overhead-constrained method leads to significant improvements in reconstructed video quality.
{"title":"Overhead-constrained rate-allocation for scalable video transmission over networks","authors":"Bo Hong, Aria Nosratinia","doi":"10.1109/DCC.2002.999998","DOIUrl":"https://doi.org/10.1109/DCC.2002.999998","url":null,"abstract":"Summary form only given. Forward error correction (FEC) based schemes are use widely to address the packet loss problem for Internet video. Given total available bandwidth, finding optimal bit allocation is very important in FEC-based video, because the FEC bit rate limits the rate available to compress video. We want to give proper protection to the source, but also prevent unwanted FEC rate expansion. The rate of packet headers is often ignored in allocating bit rate. We show that this packetization overhead has significant influence on system performance in many cases. Decreasing packet size increases the rate of packet headers, thus reducing the available rate for the source and its FEC codes. On the other hand, smaller packet size allows a larger number of packets, in which case it can be shown that the efficiency of FEC codes improves. We show that packet size should be optimized to balance the effect of packet headers and the efficiency of FEC codes. We develop a probabilistic framework for the solution of rate allocation problem in the presence of packet overhead. We implement our solution on the MPEG-4 fine granularity scalability (FGS) mode. To show the flexibility of our technique, we use an unequal error protection scheme with FGS. Experimental results show that our overhead-constrained method leads to significant improvements in reconstructed video quality.","PeriodicalId":420897,"journal":{"name":"Proceedings DCC 2002. Data Compression Conference","volume":"79 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132161557","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We consider the case of two correlated non-binary sources. Data compression is achieved by transforming the sequences of non-binary symbols into sequences of bits and then using punctured turbo codes as source encoders. Each source is compressed without knowledge about the other source, and no information about the correlation between sources is required in the encoding process. Compression is achieved because of puncturing, which is adjusted to obtain the desired compression rate. The source decoder utilizes iterative schemes over the compressed binary sequences, and recovers the non-binary symbol sequences from both sources. The performance of the proposed scheme is close to the theoretical limit predicted by the Slepian-Wolf (1973) theorem.
{"title":"Data compression of correlated non-binary sources using punctured turbo codes","authors":"Ying Zhao, J. Garcia-Frías","doi":"10.1109/DCC.2002.999962","DOIUrl":"https://doi.org/10.1109/DCC.2002.999962","url":null,"abstract":"We consider the case of two correlated non-binary sources. Data compression is achieved by transforming the sequences of non-binary symbols into sequences of bits and then using punctured turbo codes as source encoders. Each source is compressed without knowledge about the other source, and no information about the correlation between sources is required in the encoding process. Compression is achieved because of puncturing, which is adjusted to obtain the desired compression rate. The source decoder utilizes iterative schemes over the compressed binary sequences, and recovers the non-binary symbol sequences from both sources. The performance of the proposed scheme is close to the theoretical limit predicted by the Slepian-Wolf (1973) theorem.","PeriodicalId":420897,"journal":{"name":"Proceedings DCC 2002. Data Compression Conference","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114684128","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We consider compression of multi-component map images by context modeling and arithmetic coding. We apply an optimized multi-level context tree for modeling the individual binary layers. The context pixels can be located within a search area in the current layer, or in a reference layer that has already been compressed. The binary layers are compressed using an optimized processing sequence that makes maximal utilization of the inter-layer dependencies. The structure of the context tree is a static variable depth binary tree, and the context information is stored only in the leaves of the tree. The proposed technique achieves an improvement of about 25% over a static 16 pixel context template, and 15% over a similar single-level context tree.
{"title":"Context tree compression of multi-component map images","authors":"P. Kopylov, P. Fränti","doi":"10.1109/DCC.2002.999959","DOIUrl":"https://doi.org/10.1109/DCC.2002.999959","url":null,"abstract":"We consider compression of multi-component map images by context modeling and arithmetic coding. We apply an optimized multi-level context tree for modeling the individual binary layers. The context pixels can be located within a search area in the current layer, or in a reference layer that has already been compressed. The binary layers are compressed using an optimized processing sequence that makes maximal utilization of the inter-layer dependencies. The structure of the context tree is a static variable depth binary tree, and the context information is stored only in the leaves of the tree. The proposed technique achieves an improvement of about 25% over a static 16 pixel context template, and 15% over a similar single-level context tree.","PeriodicalId":420897,"journal":{"name":"Proceedings DCC 2002. Data Compression Conference","volume":"96 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123183984","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2002-04-02DOI: 10.1109/DCC.2002.1000027
Sacha Zyto, A. Grama, W. Szpankowski
Summary form only given. A wide variety of matrix transforms have been used for compression of image and video data. Transforms have also been used for motion estimation and quantization. One such transform is the singular-value decomposition (SVD) that relies on low rank approximations of the matrix for computational and storage efficiency. In this study, we describe the use of a variant of SVD in image and video compression. This variant, first proposed by Peleg and O'Leary, called semidiscrete decomposition (SDD), restricts the elements of the outer product vectors to 0/1/-1. Thus approximations of much higher rank can be stored for the same amount of storage. We demonstrate the superiority of SDD over SVD for a variety of compression schemes. We also show that DCT-based compression is still superior to SDD-based compression. We also demonstrate that SDD facilitates fast and accurate pattern matching and motion estimation; thus presenting excellent opportunities for improved compression.
{"title":"Semi-discrete matrix transforms (SDD) for image and video compression","authors":"Sacha Zyto, A. Grama, W. Szpankowski","doi":"10.1109/DCC.2002.1000027","DOIUrl":"https://doi.org/10.1109/DCC.2002.1000027","url":null,"abstract":"Summary form only given. A wide variety of matrix transforms have been used for compression of image and video data. Transforms have also been used for motion estimation and quantization. One such transform is the singular-value decomposition (SVD) that relies on low rank approximations of the matrix for computational and storage efficiency. In this study, we describe the use of a variant of SVD in image and video compression. This variant, first proposed by Peleg and O'Leary, called semidiscrete decomposition (SDD), restricts the elements of the outer product vectors to 0/1/-1. Thus approximations of much higher rank can be stored for the same amount of storage. We demonstrate the superiority of SDD over SVD for a variety of compression schemes. We also show that DCT-based compression is still superior to SDD-based compression. We also demonstrate that SDD facilitates fast and accurate pattern matching and motion estimation; thus presenting excellent opportunities for improved compression.","PeriodicalId":420897,"journal":{"name":"Proceedings DCC 2002. Data Compression Conference","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123215075","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：