Search

Still (or intra-picture) image coding includes compression of binary images and images with multiple amplitude levels, the latter case either monochrome or with multiple color components, as typically captured by cameras or image scanners. Beyond that, compression schemes are also applied to other sensor data such as magnetic resonance intensity in medical imaging, infrared, x-ray, depth captured by range sensors, or to synthetic content, specifically generated by computer graphics or captured from screens. Various methods such as run-length coding and conditional entropy coding for binary images; vector quantization, predictive coding and transform coding for multi-amplitude images are presented in more detail. Further sections discuss lossless image compression, synthesis based coding, fractal coding and 3D image coding. As examples for the overall combination of building blocks, still image compression standards like JPEG, JPEG 2000 and the intra-picture compression of AVC and HEVC video coding standards are discussed in detail.

This chapter introduces fundamental concepts of signal coding, as they are commonly applied for various types of multimedia signals (image, video, graphics, audio, speech) and associated components (e.g. depth maps of images/video). A common property is correlation between samples, which can be removed by signal processing, performing a mapping into an equivalent representation which exposes less statistical dependencies between samples and is sparse, i.e. concentrating information in few relevant samples. The main approaches for this are predictive coding, where the original signal is mapped to a prediction error signal, and transform coding, where a mapping into transform coefficients is made. Sparseness is important, as it allows representing many values by zero, which can be encoded quite efficiently. On the other hand, multimedia signals can hardly be classified as stationary random processes, they typically change local statistical properties. Therefore, adaptation of the compression methods is important, both in predictive and in transform coding. Though the methods of adaptation are often specific for a certain type of multimedia signals and are with more detail discussed in upcoming chapters, common principles are introduced here.

A transform optimization based preprocessing algorithm is presented. The method integrates combination of transforms, color component weighting and CSF filtering. Experiment results showed that proposed algorithm outperforms JPEG2000 lossy and lossless algorithm in both objective and subjective quality in wide range of compression rate.

This paper proposes a genetic-fuzzy system for speech recognition. In addition to pre-processing, with mel-cepstral coefficients, the Discrete Cosine Transform (DCT) is used to generate a two-dimensional time matrix for each pattern to be recognized. A genetic algorithm is used to optimize a Mamdani fuzzy inference system in order to obtain the best model for final recognition. The speech recognition system used in this paper was named Hybrid Method Genetic-Fuzzy Inference System for Speech Recognition (HMFE).

H.264 standard is the latest video coding standard, which uses a series of advanced coding techniques, it has a great advantage than the traditional standard in the coding efficiency,error resilience capabilities, network adaptability.This article main studies the key technologies of H.264, put forward the current problems and gives solutions,last introduces some new development and applications.

In this paper, current methods about perceptual hashing were deeply researched and the defects of them were analyzed. Principles and characters of discrete cosine transform, Watson visual model and chaotic model were introduced in details. A new method about perceptual image hashing which combined the characteristics of the three models was designed. This method use discrete cosine transform to feature extraction from image, eigenvector was handled by contrast sensitivity table, Logistic equation was used as chaos sequence generator to encrypt and at last the prediction differential method was used to quantization coding. Experimental results indicated that the method could resist the content-preserving modifications, and possessed strong robustness, safety and retrieval ability, and the collision rate decreases to level 10− 7. Therefore, this technique has the applied value in image authentication, copyright protection, image security and content-based image retrieval and so on.

Transform coding with multiple blocks results in high complexity of HEVC. Butterfly combined with multipliers method provides an efficient implementation. However, the matrix multiplication is inevitable based on this method. In this paper, matrices and element decomposition are proposed to decompose the multiplication matrices into orthogonal matrices and general matrices with smaller elements. The number of different elements in these matrices can be reduced after decomposing, so that the multiplications can be reduced. Further, the decomposed elements become smaller and much closer to power of 2, which means we can use shifting and adding to achieve the multiplications conveniently. The proposed method is suitable for the hardware implementation.

Absolute Moment Block Truncation Coding (AMBTC) is an efficient and fast lossy compression technique for still images. It is easy to implement compared to transform coding. Interpolative AMBTC is a modified method of AMBTC compression. In this paper image data hiding in images using Interpolative AMBTC is proposed. In this method the secret image is stored in the bit plane of the AMBTC compressed image and is recovered during decompression process. Experimental results show that the proposed method achieves good quality decompressed image with low computational complexity.

Still image compression algorithms have been long aiming at gray-scale image compression, in this paper we exploit a new method to improve coding efficiency of color image compression. Firstly, the combination of color space transform and wavelet transform is examined and found out the best one. Then a weighting scheme is proposed to improve objective quality of reconstruction image further. Experimental results show that the proposed scheme not only outperforms JPEG2000 lossy transform scheme and JPEG2000 lossless transform scheme in objective quality by 1–2 db. By conservative estimation, the proposed algorithm outperforms JPEG2000 lossy algorithm from 0.2BPP to 6 BPP and can be used in most practical compression situations.

The problem of stamp recognition addressed here involves a multi-stage approach which includes stamp detection, localization and segmentation, features extraction and finally, classification. In this paper we focus on the two last stages, namely features extraction by means of Point Distance Histogram and Discrete Cosine Transform, and classification employing distance calculation by means of Euclidean metrics. The first stage which leads to automatic stamps segmentation has been described in several previous papers and it is based mainly on color segmentation. The feature extractor described here works on binary images of stamps and employs polar representation of points gathered in a histogram form, which is later reduced by means of Discrete Cosine Transform. At the classification stage, compact descriptors of stamps are compared according to the distance to the reference objects (class’ centers), and the closest class is taken as the answer. The paper includes some results of selected experiments on real documents having different types of stamps. A comparison with the classical two-dimensional DCT calculated over the images is also provided to prove high discriminative power of the developed approach.

Robust Video watermarking for M-JPEG data stream based on DCT-Transform and DPCM Encoder is introduced. For this purpose the DPCM encoder is modified and feeds with DCT spectral coefficients of the incoming frames. The difference of the spectra of two sequential frames are used for the embedding the watermarks. To enhance the embedding performance and the security of watermarks some rearrangement of the pixels of the video frames and of the watermarks are introduced. The first permutation was applied to generate a random distribution of the pixels inside a block. The second was applied on the watermark and the third is responsible for the rearrangement of the frame order inside the GOP. The permutation inside the frame and the frame order increases the encryption ability and reduces visible degradation due to watermarking embedding procedure. To realize robust transmission the modified spectral coefficients are encoded into an M-JPEG data stream.

In this paper, a novel contrast enhancement technique for contrast enhancement of a low-contrast satellite image has been proposed based on the singular value decomposition (SVD) and discrete cosine transform (DCT). The singular value matrix represents the intensity information of the given image and any change on the singular values change the intensity of the input image. The proposed technique converts the image into the SVD-DCT domain and after normalizing the singular value matrix; the enhanced image is reconstructed by using inverse DCT. The visual and quantitative results suggest that the proposed SVD-DCT method clearly shows the increased efficiency and flexibility of the proposed method over the exiting methods such as the histogram equalization, gamma correction and SVD-DWT based techniques.

Among the orthogonal transforms used in video and image compression, the Discrete-Cosine-Transform (DCT) is the most commonly used one. In the existing video codecs, the motion-compensation residual (MC-residual) is transformed with the DCT. In this paper, we propose an adaptive orthogonal transform that performs better on the MC-residual than the DCT. We formulate the proposed new transform based on L1-Norm minimization with orthogonal constraints. With the DCT matrix as the starting point, it is guaranteed to derive a better orthogonal transform matrix in terms of L1-Norm minimization. The experimental results confirm that, with little side information, our method leads to higher compression efficiency for the MC-residual. Remarkably, the proposed transform performs better in the high/ complex motion situation.

In this paper, a perceptual encryption scheme based on scrambling of DCT blocks in JPEG compressed images and its variants are proposed. These schemes are suitable for the environment where perceptual degradation of the multimedia content is required. The security of these techniques is shown to be sufficient against casual attacks. These discussed schemes are kept as light as possible in order to keep the encryption overhead and cost low. Also, an investigation in the progressive degradation of image by increasing the number of the DCT blocks to be scrambled has been carried out. The quality of the degraded multimedia content is measured with objective image quality assessment (IQA) metrics such as SSIM, MS-SSIM, VIF, VIFP and UQI. These IQA metrics provide choice for the selection of control factor.

This paper introduces a multi resolution cascaded transform domain method for copyright protection of digital images. An algorithm is developed that embeds the watermark information, with out much distortion to the cover image, while allows us to extract the watermark by use of correlation. The efficiency of an image watermarking technique depends on the preservation of visually significant information. This is achieved by Dual Tree Complex Wavelet Transform (DTCWT), which is modeled to show directional selectivity and as a result consistent with the Human Visual System (HVS). The Discrete Cosine Transform (DCT) is resistant to several attacks. In this paper combined DTCWT-DCT technique is proposed, which uses a unique method for spreading, embedding and extracting the watermark. This paper compares various watermarking algorithms like DCT, Discrete Wavelet Transform (DWT), DTCWT, DCT-DWT with DTCWT-DCT method. This hybrid DTCWT-DCT shows superior performance in the presence of geometric and signal processing operations.

In this paper, a new fast transformed vector quantization (TVQ) algorithm with static pattern clustering is proposed for coding of color images. To speed up the design process of VQ with better compression ratio, the features of transform coding and VQ are combined in this work. The transformed training set is obtained using integer based orthogonal polynomials transform with reduced computational complexity. The proposed method generates a single codebook for all the three color components, because of the inter-correlation property of the proposed transformation. The proposed algorithm reduces codebook construction time by clustering the static input patterns and eliminates redundant computations at the successive iterations of codebook optimization phase of LBG codebook training process. The computation time is also reduced by choosing only non homogeneous vectors for codebook initialization from the large volume of transformed input training set. The experiments are conducted and compared with existing technique.

In the popularity of the web 2.0, transformed video copies can be found everywhere. For information retrieval, for instance, search engines can filter out videos with the same content. Also, video copy detection can protect the copyright. In this paper, we propose a fast content-based video copy detection approach by using a dynamic programming method based on discrete cosine transform to find video sequence matching. Experimental results confirm the effectiveness of the proposed approach using the large scale TRECVID2009 data set.

The aim of the paper is to explore the application of 2-D dual tree discrete wavelet transform (DDWT) which is directional and redundant over the critically sampled transform like discrete wavelet transform (DWT) for image coding. In this paper image coding application is investigated with DDWTs along with iterative projection based noise shaping (IP-NS) algorithm. IP-NS is one of sparsifying method for DDWT coefficients used to modify large coefficients to compensate for the loss of small coefficients, without substantially changing the original image. Promising results are compared with DWT and DWT with noise shaping also. After thorough investigations, it is proposed that by employing DDWT along with noise shaping algorithm significantly improve the performance over DWT.

This paper proposes a new spatial scalable and low-complexity video compression algorithm based on multiplication free three-dimensional discrete pseudo-cosine transform (3-D DPCT). Practical results which show the compression efficiency of the proposed algorithm in comparison with H.264/SVC standard are presented.

We propose a new sort-based transform for lossless data compression that can replace the BWT transform in the block-sorting data compression algorithm. The proposed transform is a parametric generalization of the BWT and the RadixZip transform proposed by Vo and Manku (VLDB, 2008), which is a rather new variation of the BWT. For a class of parameters, the transform can be performed in time linear in the data length. We give an asymptotic compression bound attained by our algorithm.

There is a major research effort under way to improve image and video coding efficiency through exploiting visual redundancy, in alignment with traditionally predictive coding and transform coding. It is motivated from the fact that natural images not only can be generally decomposed into texture and piecewise smooth parts called cartoon (e.g. edges), but may be recognized to consist of an overwhelming number of visual patterns generated by very diverse stochastic processes in nature. This paper explores perceptual non-parametric sampling methods into standardized video engine with structure-based prediction, and further suggests a learning-based framework for compressing image and video at low bit rate, by incorporating effective state-of-the-art inference algorithms to pursue an online synthesis solution. A crucial component is presented to learn the relationship (projection) between the abstracted patches (visual pattern) and the corresponding detail (feature space) in spatio-temporal manner. The experiment result shows the promising prospect for perceptual image and video coding.

In this paper, we propose a novel algorithm, named as Spatially Varying Transform (SVT). The basic idea of SVT is that we do not restrict the transform coding inside normal block boundary but adjust it to the characteristics of the prediction error. With this flexibility, we are able to achieve coding efficiency improvement by selecting and coding the best portion of the prediction error in terms of rate distortion tradeoff. The proposed algorithm is implemented and studied in the H.264/AVC framework. We show that the proposed algorithm achieves 2.64% bit-rate reduction compared to H.264/AVC on average over a wide range of test set. Gains become more significant at high bit-rates and the bit-rate reduction can be up to 10.22%, which makes the proposed algorithm very suitable for future video coding solutions focusing on high fidelity applications. The decoding complexity is expected to be decreased because only a portion of the prediction error needs to be decoded.

Compared with spatial domain based watermarking techniques, transform domain based watermarking techniques have become the main stream of this research area for a long time, since transform domain based watermarking schemes can provide not only good watermarked image quality, but also stronger robustness under general attacks or noise affection. In this chapter, our focus is shifting to the transform domain based watermarking scheme, where a watermarking scheme based on the most popular discrete cosine transform (DCT) is presented. The DCT-based scheme first transforms the cover image into frequency domain. It then selects a number of DCT bands according to the user-specified key and modifies these bands to carry the watermark bits. To have better coding results, the concept of introducing intelligent techniques into the watermarking scheme is employed again. Here a training procedure named genetic band selection (GBS) is illustrated. It employs the genetic algorithm (GA) to select a group of suitable DCT bands for watermarking. The trained result is then used in the mentioned DCT-based watermarking scheme. With the trained result of the GBS procedure, we expect the original watermarking scheme could have better performance.

In this paper, we prove that the Principal Component Analysis (PCA) and the Linear Discriminant Analysis (LDA) can be directly implemented in the DCT (Discrete Cosine Transform) domain and the results are exactly the same as the one obtained from the spatial domain. In some applications, compressed images are desirable to reduce the storage requirement. For images compressed using the DCT, e.g., in JPEG or MPEG standard, the PCA and LDA can be directly implemented in the DCT domain such that the inverse DCT transform can be skipped and the dimensionality of the original data can be initially reduced to cut down computational cost.

In this research, we propose to use the discrete cosine transform to approximate the cumulative distributions of data cube cells' values. The cosine transform is known to have a good energy compaction property and thus can approximate data distribution functions easily with small number of coefficients. The derived estimator is accurate and easy to update. We perform experiments to compare its performance with a well-known technique - the (Haar) wavelet. The experimental results show that the cosine transform performs much better than the wavelet in estimation accuracy, speed, space efficiency, and update easiness.

Designing a lossy source code remains one of the important topics in information theory, and has a lot of applications. Although plain vector quantization (VQ) can realize any fixed-length lossy source coding, it has a serious drawback in the computation cost. Companding vector quantization (CVQ) reduces the complexity by replacing vector quantization with a set of scalar quantizations. It can represent a wide class of practical VQs, while the structure in CVQ restricts it from representing every lossy source coding. In this article, we propose an optimization method for parametrized CVQ by utilizing a newly derived distortion formula. To test its validity, we applied the method especially to transform coding. We found that our trained CVQ outperforms Karhunen-Loeve transformation (KLT)-based coding not only in the case of linear mixtures of uniform sources, but also in the case of low bit-rate coding of a Gaussian source.

This article proposes a detection method to solve the problem of determining whether the audio data has been tampered with or not. In many situations, we need to authenticate audio data as an original or some tampered version. For example, we have to verify the owner of audio data once the right is debated. By using the proposed method, we can easily authenticate whether the audio data had been altered or not and thus make the said data are more reliable. In the scheme, the audio signals are transformed by DCT, the DC and low-frequency AC are hidden in the medium-frequency AC. With the inverse transformation of DCT, we can obtain the protected audio data. To authenticate the audio signals, the protected data are transformed by DCT; then the system extracts the embedded DC and low-frequency AC from the medium-frequency AC. After the inverse DCT, we can extract the audio data and compare with the protected data. The extracted audio data will greatly differ from the protected data if these data are altered; without tampering, the two are quite similar. Also, if the protected audio data are distorted by MP3 compression, the previous situation still holds.

Steganography has long been a means of secure communication. Security is achieved by camouflaging the secret message. In this paper, we present a Genetic Algorithm based Steganography using Discrete Cosine Transforms(GASDCT) and Genetic Algorithm based Steganography using DiscreteWavelets Transform(GASDWT). In our approach, the Discrete Cosine Transform and Discrete Wavelet Transform are applied to the payload. Genetic Algorithm is used to generate many stego-images based on Fitness functions; one of these which give least statistical evidence of payload is selected as the best stego image to be communicated to the destination. It is observed that GASDWT has an improvement in Bit Error Rate(BER), Peak Signal to Noise Ratio(PSNR) and embedding capacity as compared to GASDCT.

In this paper, we propose an embedded variable bit-rate (VBR) audio coder to provide the fittest quality of service (QoS) and better connectivity of service for the ubiquitous speech communications. It has scalable bandwidth for narrowband to wideband speech signal, and embedded 8 32 kbit/s VBR corresponding to the network condition and terminal capacity. For the design of the embedded VBR coder, the narrowband signals are compressed by an existing standard speech coding method for the compatibility with G.729 coder, and then the other signals are compressed hierarchically on the basis of CELP enhancement and transform coding with temporal noise shaping (TNS) method. By the objective and subjective quality tests, it is shown that the proposed embedded VBR audio coder provides a reasonable quality compared with existing audio coders such as G.722 and G.722.2 in terms of mean opinion score (MOS) and perceptual evaluation of speech quality of wideband (PESQ-WB).

In this work, we design an efficient algorithm for color image compression using a model for the rate-distortion connection. This model allows the derivation of an optimal color components transform, which can be used to transform the RGB primaries or matrices into a new color space more suitable for compression. Sub-optimal solutions are also proposed and examined. The model can also be used to derive optimal bits allocation for the transformed subbands. An iterative algorithm for the calculation of optimal quantization steps is introduced using the subband rates (entropies). We show that the rates can be approximated based on a probabilistic model for subband transform coefficients to reduce the algorithm's complexity. This is demonstrated for the Discrete Cosine Transform (DCT) as the operator for the subband transform and the Laplacian distribution assumption for its coefficients. The distortion measure considered is the MSE (Mean Square Error) with possible generalization to WMSE (Weighted MSE). Experimental results of compressed images are presented and discussed for two versions of the new compression algorithm.

Reliable methods for measuring the perceived image quality are needed to evaluate visual artifacts brought about by digital compression algorithms such as JPEG. This paper presents an objective quality-assessment method based on a circular back-propagation (CBP) neural structure: the network is trained to predict quality ratings, as scored by human assessors, from numerical features that characterize images. As such, the method aims at reproducing perceived image quality, rather than at defining a comprehensive model of the human visual system. The neural model allows one to decouple the task of feature selection from the mapping of these features into a quality score. Experimental results on a public database of test images confirm the effectiveness of the approach.

This paper presents a new approach on image-adaptive spread-transform dither modulation (STDM). The approach is performed in the discrete cosine transform (DCT) domain, and modifies the original STDM in such a way that the spread vector is weighted by a set of just noticeable differences (JND's) derived from Watson's model before it is added to the cover work. An adaptive quantization step size is next determined according to the following two constraints: 1) the covered work is perceptually acceptable, which is measured by a global perceptual distance; 2) the covered work is within the detection region. We derive the strategy on the choice of the quantization step. Further, an effective solution is proposed to deal with the amplitude scaling attack, where the scaled quantization step is produced using an extracted signal in proportion to the amplitudes of the cover work. Experimental results demonstrate that the proposed approach achieves the improved robustness and fidelity.

High-Definition (HD) videos often contain rich details as well as large homogeneous regions. To exploit such a property, Variable Block-size Transforms (VBT) should be in place so that transform block size can adapt to local activities. In this paper, we propose a 16× 16 Integer Cosine Transform (ICT) for HD video coding, which is simple and efficient. This 16×16 ICT is integrated into the AVS Zengqiang Profile and used adaptively as an alternative to the 8×8 ICT. Experimental results show that 16×16 transform can be a very efficient coding tool especially for HD video coding.

The digital television of the future will also have digital associated sound. In this way an audio quality can be achieved which is far better than that obtained with the FM transmission system used in analogue television. However, in order to keep a digital television transmission within the realms of a realistic bandwidth, a suitable bit-rate reduction of the audio source signal is required. Moreover, the transition to a digital associated sound should offer the content providers the opportunity to choose between various data rates and thus between various qualities of audio signals. Furthermore, several alternative methods (such as dual-channel audio, stereo, surround-sound etc.) should be available to the content provider. It goes without saying that certain compatibility requirements must be satisfied so that each receiver is in a position to decode the audio signal regardless of which system or bit rate has been chosen by the content provider.

In this paper we first present a brief introduction of the Sense Matrix Model (SMM), which employs a word-sense matrix representation of text for information retrieval, and then a discussion about one of the document transform techniques introduced by the model, namely discrete cosine transform (DCT) on document matrices and vectors. A first system implementation along with its experimental results is discussed, which provides marginal to medium improvements and validates the potential of DCT.

New quantization method for transform coding of speech and audio signals is proposed. The spectral coefficients obtained by the first transform are split into frequency bands, and those of each band are transformed again on a band basis, resulting in another set of coefficients for each band. Then, the efficiency of Huffman coding in two transform domains is analyzed on a band basis and a domain with better performance is selected for each band as the final quantization domain. In addition, a set of domain selection patterns with frequent occurrence is pre-defined in order to decrease the number of side-information bits for indicating the selected domains. The proposed quantization method based on the dual-domain approach is applied to ITU G.722.1 signal codec and the improvement of quantization performance for various speech and audio signals is verified.

Still image coding includes compression of binary images and multiple-amplitude level (gray scale or color) images. Substantially different methods are applied for these two cases. This chapter gives a broad overview on different methods, which are in principle combinations of methods for signal decorrelation and analysis, quantization and coding, optimized for the specific characteristics of image signals. For binary images, run-length methods and methods related to conditional entropy coding are most relevant. For multiple-amplitude image signals, vector quantization, predictive coding, transform coding and fractal coding are presented in more detail. Transform coding methods can further be clustered into block transform, filterbank and wavelet transform related methods. These methods are presented mainly for examples of luminance (gray-level) compression, as typically the chrominance components are compressed by the same techniques, but are less challenging in terms of structure and hence will allow higher compression ratios. Building blocks which are necessary to understand the principles of still image coding standards like JPEG and JPEG 2000 are discussed in detail. Further important aspects relate to the robustness of still-image compression methods in the case of transmission losses, and to content-related encoding, which allows to further improve the quality by adaptation to the content properties and signal structure. The basic methods for still image coding are also important as elements within video compression methods, which will be further discussed in chapter 13. When applied to video, still image coding is also denoted as intra- frame coding, expressing that compression of a sequence of video frames is performed without exploiting the interframe redundancies.

Many video coding schemes, including MPEG-2, use a Discrete Cosine Transform (DCT). The recently completed video coding standard, H.264/AVC, uses an integer transform, which will be referred to as HT in this paper. We propose an efficient method to convert DCT coefficients to HT coefficients entirely in the transform domain. We show that the conversion is essentially a 2D transform. We derive the transform kernel matrix, provide a fast algorithm and an integer approximation of the transform. We show that the proposed transform domain conversion outperforms the conventional pixel domain approach. It is expected to have applications in transform domain video transcoding.

An image coding scheme which combines transform coding with a human visual system (HVS) model has been developed. The system include an eye tracker to pick up the point of regard of a single viewer. One can then utilize that the acuity of the HVS is lower in the peripheral vision than in the central part of the visual field. A model of the decreasing acuity of the HVS which can be applied to a wide class of transform coders is described. Such a coding system has a large potential for data compression.In this paper we have incorporated the model into an image coder based on the discrete wavelet packet transform (DWPT) scheme.

An image coding scheme which combines transform coding with a human visual system (HVS) model has been developed. The system include an eye tracker to pick up the point of regard of a single viewer. One can then utilize that the acuity of the HVS is lower in the peripheral vision than in the central part of the visual field. A model of the decreasing acuity of the HVS which can be applied to a wide class of transform coders is described. Such a coding system has a large potential for data compression.In this paper we have incorporated the model into an image coder based on the discrete wavelet transform (DWT) scheme.

In this paper, we present a novel design of a wavelet-based video coding algorithm within a conventional hybrid framework of temporal motion-compensated prediction and transform coding. Our proposed algorithm involves the incorporation of multi-frame motion compensation as an effective means of improving the quality of the temporal prediction. In addition, we follow the rate-distortion optimizing strategy of using a Lagrangian cost function to discriminate between different decisions in the video encoding process. Finally, we demonstrate that context-based adaptive arithmetic coding is a key element for fast adaptation and high coding efficiency. The combination of overlapped block motion compensation and frame-based transform coding enables blocking-artifact free and hence subjectively more pleasing video. In comparison with a highly optimized MPEG-4 (Version 2) coder, our proposed scheme provides significant performance gains in objective quality of 2.0- 3.5 dB PSNR.

A new approach for modeling audio signal will be presented, in view of efficient encoding. The method is based upon hybrid models featuring transient, tonal and stochastic components in the signal. The three components are estimated and encoded independently using a strategy very much in the spirit of transform coding. The signal models involve nonlinear expansions on local trigonometric bases, and binary trees of wavelet coefficients. Unlike several existing approaches, the method does not rely on any prior segmentation of the signal. The talk is based on joint works with L. Daudet and S. Molla.

In this correspondence, we propose a new Cellular Automata based transform coding scheme for grey level and color still images. This new scheme is markedly superior to the currently available DCT based schemes both in terms of Compression Ratio and Reconstructed Image Fidelity.

A new catastrophicity test for convolutional encoders whose rate and generator polynomials vary with time is presented. Based on this test computationally efficient algorithm to determine whether or not a time-varying convolutional encoder is catastrophic is derived. This algorithm is shown to be simpler than the catastrophicity test proposed by Balakirsky [1]. Furthermore, the algorithm can easily be generalised to rate k/n time-varying convolutional encoders.