Search

In this paper, we propose a new interpretation of the information maximization method (InfoMax) from a perspective of the rate distortion theory. We show that under specific conditions, InfoMax is equivalent to the minimization of a compression rate under the constraint of zero distortion. Zero distortion, or equivalently, zero reconstruction error between the input and its reconstruction, does not provide meaningful solutions in many cases. Based on the new interpretation, we extend InfoMax to be able to deal with non-zero distortion and also to learn under/over-complete representations. Experimental results on synthetic as well as real data show the effectiveness of our method.

Deep belief networks (DBNs) are currently the dominant technique for modeling the architectural depth of brain, and can be trained efficiently in a greedy layer-wise unsupervised learning manner. However, DBNs without a narrow hidden bottleneck typically produce redundant, continuous-valued codes and unstructured weight patterns. Taking inspiration from rate distortion (RD) theory, which encodes original data using as few bits as possible, we introduce in this paper a variant of DBN, referred to as sparse-response DBN (SR-DBN). In this approach, Kullback–Leibler divergence between the distribution of data and the equilibrium distribution defined by the building block of DBN is considered as a distortion function, and the sparse response regularization induced by L 1 -norm of codes is used to achieve a small code rate. Several experiments by extracting features from different scale image datasets show that our approach SR-DBN learns codes with small rate, extracts features at multiple levels of abstraction mimicking computations in the cortical hierarchy, and obtains more discriminative representation than PCA and several basic algorithms of DBNs.

In order to solve the problem that the general image compression algorithm and the running speed based on single DSP or PC cannot satisfied the demand for high-resolution mapping real-time image compression, this paper proposed a low-complexity rata pre-allocated algorithm and an implement mode. Firstly, the multi-level integer wavelet applied to image. Secondly, depend on the principle that various subband important to image reconstruction, and the optimal rate-distortion theory, pre-allocated the actual contribution rate in the coding for the each sub-band. Lastly, when each sub-band has been coded, the coding have been truncated in real-time according to the pre-allocated rate, so the bit rate size of the compressed image has been controlled precisely. Test results show that this paper proposed algorithm which has more features than the other classical standard compression algorithm. The features include less calculation, less usage of storage space, and simple structure of the bit rate stream. The algorithm with low complexity but the same compression performance as the JPEG2000 standard is significantly better than SPIHT and other algorithms. Finally, the algorithm proposed in this paper has been realizes in real-time by many DSP and FPGA. (C) 2012 Elsevier GmbH. All rights reserved.

In this paper, we provide an analysis of the requantization problem in order to improve the requantization process. This analysis is based on theoretical R-D results of requantized Laplacian sources instead of minimizing requantization errors as commonly found in the literature. We derive the effective quantizer characteristic by applying superposition to the quantizer characteristics of encoder and transcoder. Further investigation shows that the effective quantizer has a periodic property. Using the memoryless property of the probability distribution function and the periodic property of the effective quantizer characteristic, we derive expressions for entropy and distortion. Based on the theoretical R-D model, requantization for fine and coarse quantized signals is investigated. The analysis of the R-D behavior shows that a heuristic can be derived which improves the requantization process. Finally, the results from the R-D analysis are verified for requantization transcoding of H.264/AVC video streams. We show that the transcoding process for H.264/AVC video streams, which corresponds to coarse quantization, is improved with gains up to 1dB.

Matching pursuit decompositions have been employed for signal coding. For this purpose, matching pursuit coefficients need to be quantized. However, their behavior has been shown to be chaotic in some cases; posing difficulties to their modeling and quantizer design. In this work, a different approach is presented. Instead of trying to model the statistics of matching pursuit coefficients, the statistics of the angle between the residue signal and the element selected in each iteration of the matching pursuit are studied, what allows to model matching pursuits coefficients indirectly. This approach results in a simple statistical model. This is so because one observes that the statistics of such angles do not vary substantially after the first matching pursuit iteration, and can be approximately modeled as independent and identically distributed. Moreover, it is also observed that the probability density functions of matching pursuit angles are reasonably modeled by a single probability density function. This function depends only on the dictionary employed and not on the signal source. The derived statistical model is validated by employing it to design Lloyd-Max quantizers for matching pursuit coefficients. The Lloyd-Max quantizers obtained show good ratexdistortion performance when compared to the state-of-the-art methods.

In this paper, we propose three novel multiple description coding (MDC) methods for reliable transmission of compressed animated meshes represented by series of 3D static meshes with same connectivity. The proposed methods trade off reconstruction quality for error resilience to provide the best expected reconstruction of 3D mesh sequence at the decoder side. The methods are based on layer duplication and partitioning of the set of vertices of a scalable coded animated mesh by either spatial or temporal subsampling. Each partitioned set is encoded separately to generate independently decodable bitstreams or so-called descriptions. In addition, the proposed MDC methods can achieve varying redundancy allocations by including a number of encoded spatial or temporal layers from the other description. The algorithms are evaluated with redundancy-rate-distortion (RRD) curves and per-frame reconstruction analysis. RRD performances show that vertex partitioning-based MDC performs better at low redundancies for especially spatially dense models. Temporal subsampling-based MDC performs better at moderate redundancies as well as low redundancies for spatially coarse models. Layer duplication-based MDC can achieve the lowest redundancies with flexible redundancy allocation capability and can be designed to achieve the smallest variance of reconstruction quality between consecutive frames.

There are abundant intra and inter prediction modes in the AVS video coding standard. Rate distortion optimized mode decision can fully utilize this flexibility to improve the spatio-temporal prediction efficiency and maximize the coding efficiency. However, the implementation complexity is dramatically high due to huge throughput burden. Hardware oriented mode decision algorithm is tailored for VLSI implementation in this work for high definition video coding. Mode preselection is employed to alleviate the dramatic throughout burden. Also, intelligent pipeline scheduling mechanism is proposed to break the intrinsic data dependency in intra prediction, which is directly related with mode decision. The proposed simplified algorithm is well-suited for hardware implementation with small performance penalty. Finally, the VLSI architecture is proposed with good trade off between circuit consumption and rate distortion performance.

In this paper, a reversible integer to integer time domain lapped transform (RTDLT) is introduced. TDLT can be taken as a combination of time domain pre- and post-filter modules with discrete cosine transform (DCT). Different from TDLT, the filters and DCT in our proposed RTDLT are realized from integer to integer by multi-lifting implementations after factorizing the filtering and transforming matrixes into triangular elementary reversible matrices (TERMs). Lifting implementations are realized by only shift and addition without any floating-point multiplier to reduce complexity. The proposed method can realize progressive lossy-to-lossless image compression with a single bit-stream. Simulation results show that RTDLT-based compression system obtains comparable or even higher compression-ratio in lossless compression than that of JPEG2000 and JPEG-LS, as well as gratifying rate distortion performance in lossy compression. Besides, RTDLT keeps low-complexity in hardware realization because it can be parallel implemented on the block level.

In this paper, a redundant picture formation algorithm that takes into account a given redundancy rate constraint is presented for error resilient wireless video transmission without reliance on retransmissions. The algorithm assigns priorities to macroblocks (MBs) according to two suggested metrics and ranks MBs accordingly. The first metric is based on an end-to-end distortion model and aims at maximising the reduction in distortion per redundancy bit. The end-to-end distortion accounts for the effects of error propagation, mismatch between the primary and redundancy description and error concealment. MBs providing large distortion reduction for fewer bits spent are assigned a higher priority. The second metric employs the variance of the motion vectors of a macroblock and those of its neighbouring blocks. Results show that the rate distortion metric outperforms other examined metrics by up to 2dB. A comparison with other error resilience schemes, such as loss adaptive-rate distortion optimisation (LA-RDO) and cross packet forward error correction (FEC), is also presented. Results show that our approach outperforms the LA-RDO solution. Cross packet FEC provides better performance in the presence of random packet losses. However, results show that our approach can tackle bursty losses better. With bursty errors, the correction capabilities of the Reed-Solomon codes can be exceeded, and the availability of a coarse representation of the coded macroblocks, through a redundant slice scheme, can offer performance gains of up to 1dB.

This paper proposes a novel compression scheme with a tunable complexity-rate-distortion trade-off. As images increase in size and resolution, more efficient compression schemes with low complexity are required on-board Earth observation satellites. The standard of the Consultative Committee for Space Data Systems (CCSDS) defines a strip-based compression scheme with the advantages of a low complexity and an easy rate control [CCSDS, Image Data Compression Recommended Standard CCSDS 122.0-B-1 Blue Book, November 2005]. However, future mission specifications expect higher performance in terms of rate-distortion. The scheme proposed in this paper intends to perform better than the CCSDS standard while preserving low complexity and easy rate control. Moreover, to comply with existing on-board devices, the proposed core compression engine still uses the wavelet transform but in association with a linear post-processing inspired from the bandelet transform. The post-transform decomposes a small block of wavelet coefficients on a particular basis. This basis is adaptively selected within a predefined dictionary by rate-distortion optimization. The computational complexity depends upon the dictionary size and of the basis structure. An extremely simple dictionary, reduced to the Hadamard basis, is proposed. The post-transform efficiency is illustrated by experiments on various Earth observation images provided by the French Space Agency (CNES).

In the joint multiview video model (JMVM) proposed by JVT, the variable block-size motion estimation (ME) and disparity estimation (DE) have been employed to determine the best coding mode for each macroblock (MB). These give a high coding efficiency for multiview video coding (MVC), however, they cause a very high computational complexity in encoding system. This paper proposes to reduce the complexities of the ME and DE processes with an early SKIP mode decision algorithm based on the analysis of prediction mode distribution regarding the corresponding MBs in the neighbor view. In this method, the mode decision procedures of most of MBs can be early terminated, and thus much of computation for ME and DE can be greatly reduced. Simulation results demonstrate that our algorithm can achieve computational saving of 46-57% (depending on the tested sequences) with no significant loss of rate-distortion performance.

Wyner-Ziv coding enables low complexity video encoding with the motion estimation procedure shifted to the decoder. However, the accuracy of decoder motion estimation is often low, due to the absence of the input source frame (at the decoder). In this paper, we propose a novel Wyner-Ziv successive refinement approach to improve the motion compensation accuracy and the overall compression efficiency of Wyner-Ziv video coding. Our approach encodes each frame by multiple Wyner-Ziv coding layers and uses the progressively refined reconstruction frame to guide the motion estimation for progressively improved accuracy. The proposed approach yields competitive results against state-of-the-art low complexity Wyner-Ziv video coding approaches, and can gain up to 3.8dB over the conventional Wyner-Ziv video coding approach and up to 1.5dB over the previous bitplane-based refinement approach. Furthermore, this paper also presents the rate distortion analysis and the performance comparison of the proposed approach and conventional approaches. The rate distortion performance loss (due to performing decoder motion estimation) is at most 2.17dB (or equivalently 14nats/pixel) in our scheme according to our analysis, but can be more than 6dB in the conventional approach according to previous research. For the simplified two-layers case of our approach, we derive the optimal subsampling ratio in the sense of rate distortion performance. We also extend our analysis and conclusions from P frame to B frame. Finally, we verify our analysis by experimental results.

In this paper, we first develop a simple but nevertheless accurate and fully analytical model for the sequence-level distortion-rate (D-R) performance of predictive video source encoding. This model focuses on the input-output behavior of the source encoder and requires very limited amount of empirical data to set up its working parameters, namely three pairs of rate and distortion. Then, we propose an accurate and fully analytical model for the distortion due to lost frames in wireless communication systems. The proposed model builds on long-term average properties of the video source. In our analysis, we combine the properties of the video sequence and those of the wireless transmission link. Finally, we provide an asymptotic analysis of the video transmission through a wireless channel. We compare the results to those obtained by using less sophisticated models, like a (D-R)-function that is built on a Gaussian memoryless source. It is shown that the obtained results are substantially different.

This paper addresses video transcoding from H.264/AVC into MPEG-2 with reduced complexity and high rate-distortion efficiency. While the overall concept is based on a cascaded decoder-encoder, the novel adaptation methods developed in this work have the advantage of providing very good performance in H.264/AVC to MPEG-2 transcoding. The proposed approach exploits the similarities between the coding tools used in both standards, with the objective of obtaining a computationally efficient transcoder without penalising the signal quality. Fast and efficient methods are devised for conversion of macroblock coding modes and translation of motion information in order to compute the MPEG-2 coding format with a reduced number of operations, by reusing the corresponding data embedded in the incoming H.264/AVC coded stream. In comparison with a cascaded decoder-encoder, the fast transcoder achieves computational complexity savings up to 60% with slightly better peak signal-to-noise ratio (PSNR) at the same bitrate.

Layered compression paradigms such as scalable, multiple description, and multi-view video coding, require coupled encoding decisions among layers to achieve optimal distortion performance under buffer constraints. Moreover, due to the dynamic and time-varying source characteristics, and temporal coupling of encoding decisions through the buffer constraints, it is not only necessary to consider the immediate rate-distortion impact of encoding decisions, but also their long-term rate-distortion impact. In other words, optimal encoding decisions must consider the coupling between layers and the coupling across time. In many scenarios, however, it may be impractical to make joint coding decisions for all of the layers. For instance, a two layer bitstream may be coded using different encoders for the base layer and enhancement layer, each with its own autonomous control plane; or, if the same encoder is used for multiple layers, then a joint decision process, which considers the aforementioned dependencies, may be too complex. In this paper, we propose a framework for autonomous decision making in layered video coders, which decouples the decision making processes at the various layers using a novel layered Markov decision process. We illustrate how this framework can be applied to decompose the decision processes for several typical layered video coders with different dependency structures and we observe that the performance of the proposed decomposition highly depends on the ability of the layers to model each other.

Variational principles for the rate distortion (RD) theory in lossy compression are formulated within the ambit of the generalized nonextensive statistics of Tsallis, for values of the nonextensivity parameter satisfying $ 0 < q < 1 $ and $ q > 1 $. Alternating minimization numerical schemes to evaluate the nonextensive RD function, are derived. Numerical simulations demonstrate the efficacy of generalized statistics RD models., 31 pages, 2 figures. Minor formatting and typographical changes

In this paper, we propose a novel two-pass rate control algorithm to achieve constant quality for H.264/AVC high definition video coding. With the first-pass collected rate and distortion information and the built model of scene complexity, the encoder can determine the expected distortion which could be achieved in the second-pass encoding under the target bit rate. According to the built linear distortion-quantizer (D-Q) model, before encoding one frame, the quantization parameter can be solved to realize constant quality encoding. After encoding one frame, the model parameters will be updated with linear regression method to ensure the prediction accuracy of the quantization parameter of next encoded frame with the same coding type. In order to obtain the expected distortion of each frame under the target bit rate, a GOP-level bit allocation scheme is also designed to adjust the target bit rate of each GOP based on the scene complexity of the GOP in the second-pass encoding. In addition, the effect of scene change on the updating of D-Q model is considered. The model will be re-initialized at the scene change to minimize modeling error. The experimental results show that compared with the latest two-pass rate control algorithm, our proposed algorithm can significantly improve the bit control accuracy at comparable coding performance in terms of constant quality and average PSNR. On average, the improvement of bit control accuracy achieved about 90%.

Although it was proven in the 1970s already by Wyner and Ziv and Slepian and Wolf that, under certain conditions, the same rate-distortion boundaries exist for distributed video coding (DVC) systems as for traditional predicting systems, until now no practical DVC system has been developed that even comes close to the performance of state-of-the-art video codecs such as H.264/AVC in terms of rate-distortion. Some important factors for this are the lower accuracy of the motion estimation performed at the decoder, the inaccurate modeling of the correlation between the side information and the original frame, and the absence in most state-of-the-art DVC systems of anything conceptually similar to the notion of skipped macroblocks in predictive coding systems. This paper proposes an extension of a state-of-the-art transform domain residual DVC system with an implementation of skip mode. The skip mode has an impact at two different places: in the turbo decoder, more specifically the soft input, soft output (SISO) convolutional decoder, and in the puncturing of the parity bits. Results show average bitrate gains up to 39% (depending on the sequence) achieved by combining both approaches. Furthermore, a hybrid video codec is presented where the motion estimation task is shifted back to the encoder. This results in a drastic increase in encoder complexity, but also in a drastic performance gain in terms of rate-distortion, with average bitrate savings up to 60% relative to the distributed video codec. In the hybrid video codec, smaller but still important average bitrate gains are achieved by implementing skip mode: up to 24%.

Multiple description lattice vector quantization (MDLVQ) is an effective technique to realize multiple description coding. In this paper, we investigate a 3-description lattice vector quantization (LVQ) design using A"2 lattice, which finds a 3-tuple combination of sublattice points to represent each fine lattice point, based on two principles that each edge of the triangle formed by the 3-tuple points needs to be as short as possible, and the gravity center of the triangle is as close as possible to the fine lattice point. Following a delicate sublattice partition, a well-designed construction and mapping of each fine lattice point to a 3-tuple of sublattice points is developed to minimize the side distortion. The proposed index assignment is shown to achieve smaller distortion (up to 0.58dB for Gaussian source) than some other 3-description index assignments. Compared with a 2-description LVQ, the 3-description LVQ achieves better expected rate-distortion performance in most cases of high description loss rates, while it also exhibits more graceful coding results with significantly smaller side distortions. Numerical analysis on Gaussian source and the simulations on image source validate the effectiveness of our proposed index assignment scheme.

The H.264/AVC is the newest video coding standard recommended by ITU-T and MPEG. Compared with all existing video coding standards, H.264 can achieve superior performance by using many advanced techniques. Intra mode selection is an important feature in H.264 standard and can reduce the spatial redundancy in intra-frame significantly. An efficient rate distortion optimization (RDO) technique is employed in H.264 to choose the best mode for each MB, but the computational cost increases drastically. In this paper, a fast intra-mode selection algorithm is introduced. By using a fast edge detection method which is based on non-normalized Haar transform (NHT), edge for each sub-block can be extracted. Based on the local edge information, only few intra-modes are chosen as mode candidates. A fast RDO algorithm is also proposed in this paper which is based on accurate rate-distortion estimation model and the fast intra-mode RDO method. By combining these two methods, computational load is reduced remarkably. Experimental results show that this fast intra-mode selection scheme can lessen the encoding time significantly with little loss of bit-rate and visual quality.

In this paper, an H.264/AVC-based multiple description video coding scheme is proposed. It utilizes the advanced video coding tools and features provided in H.264/AVC to introduce redundancy into descriptions. Two independently decodable descriptions are generated, each consisting of two slice groups. One of them, called main slice group (MSG), is encoded normally as main information. The other one, called side slice group (SSG), is encoded with fewer bits as redundancy information by using larger quantization step sizes. Spatial and temporal correlations between neighboring macroblocks in video frames are exploited to achieve efficient redundancy coding. Experimental results show that the proposed MDC scheme is superior to previous slice group based multiple description coding (MDC) schemes in terms of the rate-distortion (R-D) performance.

Rate control plays an essential role in video coding and transmission to provide the best video quality at the receiver end given the constraint of certain network conditions. This paper proposes a rate control algorithm for the wavelet-based open-source Dirac video codec. The existing Dirac architecture has a constant-quality control mechanism based on rate-distortion optimization (RDO), giving variable bitrate. The proposed algorithm exploits the existing constant-quality control, which is governed by a parameter called quality factor (QF) to give a constant bitrate. A mathematical model called the rate-quality factor (R-QF) is derived to generate optimum QF for the current coding frame using the bitrate resulting from the encoding of the previous frame in order to meet the target bitrate. The proposed algorithm is a complete one-pass process and does not require complex mathematical computation. The process of calculating the QF is simple and, further, calculation is not required for each coded frame. It also provides the rate control solution for both intra-frame-only and inter-frame coding modes. The experimental results show that the proposed algorithm can control the bitrate precisely (i.e. within 1% of target bitrate in average for inter-frame coding mode and near-perfect flat response in generated bits vs. frame number curve in intra-frame-only coding mode). Moreover, the variation of bitrate over each group of pictures (GOP) in inter-frame coding mode is lower than that of H.264 using JM11. This is an advantage in preventing the buffer overflow and underflow for real-time multimedia data streaming. More importantly, there is no PSNR performance loss because of application of the proposed rate control algorithm. It gives superior quality over relatively static motion sequences and fast motion sequences with average quality as shown in the analysis and evaluations presented in this paper.

Bezier curves (BC) are important tools in a wide range of diverse and challenging applications, from computer-aided design to generic object shape descriptors. A major constraint of the classical BC is that only global information concerning control points (CP) is considered, consequently there may be a sizeable gap between the BC and its control polygon (CtrlPoly), leading to a large distortion in shape representation. While BC variants like degree elevation, composite BC and refinement and subdivision narrow this gap, they increase the number of CP and thereby both the required bit-rate and computational complexity. In addition, while quasi-Bezier curves (QBC) close the gap without increasing the number of CP, they reduce the underlying distortion by only a fixed amount. This paper presents a novel contribution to BC theory, with the introduction of a dynamic Bezier curve (DBC) model, which embeds variable localised CP information into the inherently global Bezier framework, by strategically moving BC points towards the CtrlPoly. A shifting parameter (SP) is defined that enables curves lying within the region between the BC and CtrlPoly to be generated, with no commensurate increase in CP. DBC provides a flexible rate-distortion (RD) criterion for shape coding applications, with a theoretical model for determining the optimal SP value for any admissible distortion being formulated. Crucially DBC retains core properties of the classical BC, including the convex hull and affine invariance, and can be seamlessly integrated into both the vertex-based shape coding and shape descriptor frameworks to improve their RD performance. DBC has been empirically tested upon a number of natural and synthetically shaped objects, with qualitative and quantitative results confirming its consistently superior shape approximation performance, compared with the classical BC, QBC and other established BC-based shape descriptor techniques.

One of the new features in the H.264/AVC encoder is the use of Lagrangian rate-distortion optimization (RDO) method during mode decision at the macroblock level. The RDO technique has been employed in H.264/AVC for intra-prediction mode selection to achieve better coding efficiency. But the computational complexity of mode decision algorithm is extremely high. To reduce the complexity of mode decision, we propose an efficient and fast 4x4 intra-prediction mode selection scheme. The proposed method reduces the candidate of the prediction modes based on the correlation between neighboring blocks and the sum of absolute transformed difference (SATD) between the original block and the intra-predicted block. Firstly, the rank of each mode is obtained based on the SATD value. Then, the candidate modes are further reduced by using the combination of rank and most probable mode. The proposed method reduces the number of candidate mode to either one or two. Simulation results demonstrate that the proposed mode decision method reduces about 91% of mode decision time and 70% of total encoding time of intra-coding with ignorable degradation of coding performance.

Due to its excellent rate-distortion performance, set partitioning in hierarchical trees (SPIHT) has become the state-of-the-art algorithm for image compression. However, the algorithm does not fully provide the desired features of progressive transmission, spatial scalability and optimal visual quality, at very low bit rate coding. Furthermore, the use of three linked lists for recording the coordinates of wavelet coefficients and tree sets during the coding process becomes the bottleneck of a fast implementation of the SPIHT. In this paper, we propose a listless modified SPIHT (LMSPIHT) approach, which is a fast and low memory image coding algorithm based on the lifting wavelet transform. The LMSPIHT jointly considers the advantages of progressive transmission, spatial scalability, and incorporates human visual system (HVS) characteristics in the coding scheme; thus it outperforms the traditional SPIHT algorithm at low bit rate coding. Compared with the SPIHT algorithm, LMSPIHT provides a better compression performance and a superior perceptual performance with low coding complexity. The compression efficiency of LMSPIHT comes from three aspects. The lifting scheme lowers the number of arithmetic operations of the wavelet transform. Moreover, a significance reordering of the modified SPIHT ensures that it codes more significant information belonging to the lower frequency bands earlier in the bit stream than that of the SPIHT to better exploit the energy compaction of the wavelet coefficients. HVS characteristics are employed to improve the perceptual quality of the compressed image by placing more coding artifacts in the less visually significant regions of the image. Finally, a listless implementation structure further reduces the amount of memory and improves the speed of compression by more than 51% for a 512x512 image, as compared with that of the SPIHT algorithm.

Region-of-interest (ROI) is an essential task that one must undertake in low bit-rate multimedia communications because of the limited bandwidth of the channels and the transcoder between different standards. In this paper, an effective ROI determination method is proposed. The skin-color extraction is first employed to determine the ROI macroblocks; then, the framework for the adjustment of the quantization parameters (QPs) of ROI macroblocks according to the distortion and bit-rate variations is proposed to fit the target bit rate. In view of the residual distortion information, the fuzzy logic controller proposed in our framework can adaptively adjust the weighting factor for corresponding QPs by the distortion variation in the macroblock layer. A linear prediction formula, derived from the rate variation, is proposed to allocate appropriate bits for each ROI macroblock and maintain the target bit rate and buffer fullness. Experiment results on TMN8 reveal that the proposed ROI video coding can significantly enhance the quality at ROIs. Furthermore, the proposed framework obtains about 0.5dB gain in the objective performance and also provides the better subjective quality compared with previous works.

In order to achieve a high compression ratio, the H.264/AVC standard has incorporated a large number of coding modes which must be evaluated during the coding process to determine the optimal rate-distortion tradeoff. The coding gains of H.264/AVC arise at the expense of significant coder complexity which may not be desired for mobile devices with limited battery life. One coder process that has been identified as having potential for achieving computation savings is the selection between skipping the coding of a macroblock and coding of the macroblock in one of the remaining coding modes. In low-motion subsequences, a large percentage of macroblocks are ''skipped'', that is, no coded data are transmitted for these macroblocks. By estimating which macroblocks are to be skipped during the coding process, significant savings in computation can be realized, since the coder then does not evaluate the rate-distortion costs of all candidate coding modes. In this work, we place this skip versus code decision in a Bayesian framework. We use the rate-distortion cost difference between coding and skipping a macroblock as the single decision feature and determine an appropriate decision threshold following modeling of the cost difference's class-conditional PDFs. Finally, in order to further limit system complexity, we model the threshold's parameters as functions of application- and sequence-specific characteristics, namely, the quantization parameter and an activity factor. This results in a decision threshold that is only a function of these two characteristics, which are either known or easily measured. It is shown that this approach can result in a time savings of over 80% for low-motion sequences at a negligible decrease or, in certain cases, a slight increase in quality over a reference H.264 codec.

Multistage vector quantization (MSVQ) and their variants have been recently proposed. Before MSVQ is designed, the user must artificially determine the number of codewords in each VQ stage. However, the users usually have no idea regarding the number of codewords in each VQ stage, and thus doubt whether the resulting MSVQ is optimal. This paper proposes the genetic design (GD) algorithm to design the MSVQ. The GD algorithm can automatically find the number of codewords to optimize each VQ stage according to the rate-distortion performance. Thus, the MSVQ based on the GD algorithm, namely MSVQ(GD), is proposed here. Furthermore, using a sharing codebook (SC) can further reduce the storage size of MSVQ. Combining numerous similar codewords in the VQ stages of MSVQ produces the codewords of the sharing codebook. This paper proposes the genetic merge (GM) algorithm to design the SC of MSVQ. Therefore, the constrained-storage MSVQ using a SC, namely CSMSVQ, is proposed and outperforms other MSVQs in the experiments presented here.

The genetic code maps the sixty-four nucleotide triplets (codons) to twenty amino-acids. Some argue that the specific form of the code with its twenty amino-acids might be a 'frozen accident' because of the overwhelming effects of any further change. Others see it as a consequence of primordial biochemical pathways and their evolution. Here we examine a scenario in which evolution drives the emergence of a genetic code by selecting for an amino-acid map that minimizes the impact of errors. We treat the stochastic mapping of codons to amino-acids as a noisy information channel with a natural fitness measure. Organisms compete by the fitness of their codes and, as a result, a genetic code emerges at a supercritical transition in the noisy channel, when the mapping of codons to amino-acids becomes nonrandom. At the phase transition, a small expansion is valid and the emergent code is governed by smooth modes of the Laplacian of errors. These modes are in turn governed by the topology of the error-graph, in which codons are connected if they are likely to be confused. This topology sets an upper bound - which is related to the classical map-coloring problem - on the number of possible amino-acids. The suggested scenario is generic and may describe a mechanism for the formation of other error-prone biological codes, such as the recognition of DNA sites by proteins in the transcription regulatory network., Comment: Keywords: genetic code, rate-distortion theory, biological information channels

Due to power consumption restrictions, low-power H.264 encoders cannot take advantage of the variable block sizes available in H.264 motion estimation. This work presents two methods to determine a block-size partition without an initial search. With both of these methods, computationally burdensome Lagrange optimization is not required. The methods are derived from a cellular nonlinear network (CNN) segmentation algorithm and, along with the partition, indicate early termination of motion estimation and the skip modes of H.264. Both methods achieve better rate-distortion performance when compared to motion estimation with only 16x16 sized blocks. The 16x16 only case is descriptive of a low-power case where the variable block sizes cannot be used. For low bitrates, both methods achieve equivalent performance when compared to Lagrange optimization. Also presented are the computational complexity of the methods and the power consumption when implemented with existing CNN hardware.

We propose a novel framework for fast mode decision in the simple and main profiles of the H264 video coding standard. Our framework consists of a specific combination of algorithms, each achieving computational savings while retaining Rate Distortion (RD) performance very similar to the standard. In particular, we utilise a set of skip mode conditions for P and B slices, two heuristics that reduce the cardinality of the inter mode set to be examined, inter/intra mode prediction and the monotonicity property of the Rate Distortion cost functions. We achieve content dependent savings in run times between 5.8 and 90.1% as compared to H264. Compared to other work that was used as input to the standard, our scheme is faster by 9-23% for very similar RD performance. The proposed framework can be used wholly or partially for computational speed-ups, it is independent of the motion search method used and is applicable in both the rate controlled and non rate controlled cases.

The demand for video access services through wireless networks, as important parts of larger heterogeneous networks, is constantly increasing. To cope with this demand, flexible compression technology to enable optimum coding performance, especially at low bit-rates, is required. In this context, scalable video coding emerges as the most promising technology. A critical problem in wavelet-based scalable video coding is bit-stream allocation at any bit-rate and in particular when low bit-rates are targeted. In this paper two methods for bit-stream allocation based on the concept of fractional bit-planes are reported. The first method assumes that minimum rate-distortion (R-D) slope of the same fractional bit-plane within the same bit-plane across different subbands is higher than or equal to the maximum R-D slope of the next fractional bit-plane. This method is characterised by a very low complexity since no distortion evaluation is required. Contrasting this approach, in the second method the distortion caused by quantisation of the wavelet coefficients is considered. Here, a simple yet effective statistical distortion model that is used for estimation of R-D slopes for each fractional bit-plane is derived. Three different strategies are derived from this method. In the first one it is assumed that the used wavelet is nearly orthogonal, i.e. the distortion in the transform domain is treated as being equivalent to the distortion in the signal domain. To reduce the error caused by direct distortion evaluation in the wavelet domain, the weighting factors are applied to the used statistical distortion model in the second strategy. In the last strategy, the derived statistical model is used during the bit-plane encoding to determine optimal position of the fractional bit-plane corresponding to refinement information in the compressed bit-stream. Results of selected experiments measuring peak signal to noise ratio (PSNR) of decoded video at various bit-rates are reported. Additionally, the PSNR of decoded video at various bit-rates is measured for two specific cases: when the methods for bit-stream allocation are used to assign quality layers in the compressed bit-stream, and when quality layers are not assigned.

In this paper, we review the recent advances in rate control techniques for video coding. The rate control algorithms recommended in the video coding standards are briefly described and analyzed. Recent advances, such as new concepts in rate-distortion modelling and quality constrained control, are presented. With these techniques, the rate control performance can be improved. The paper not only summarizes these recent rate control techniques but also provides explicit directions for future rate control algorithm design.

This paper presents a novel scheme for simultaneous compression and denoising of images: WISDOW-Comp (Wavelet based Image and Signal Denoising via Overlapping Waves-Compression). It is based on the atomic representation of wavelet details employed in WISDOW for image denoising. However, atoms can be also used for achieving compression. In particular, the core of WISDOW-Comp consists of recovering wavelet details, i.e. atoms, by exploiting wavelet low frequency information. Therefore, just the approximation band and significance map of atoms absolute maxima have to be encoded and sent to the decoder for recovering a cleaner as well as compressed version of the image under study. Experimental results show that WISDOW-Comp outperforms the state of the art of compression based denoisers in terms of both rate and distortion. Some technical devices will also be investigated for further improving its performances.

We extend high-rate quantization theory to Wyner-Ziv coding, i.e., lossy source coding with side information at the decoder. Ideal Slepian-Wolf coders are assumed, thus rates are conditional entropies of quantization indices given the side information. This theory is applied to the analysis of orthonormal block transforms for Wyner-Ziv coding. A formula for the optimal rate allocation and an approximation to the optimal transform are derived. The case of noisy high-rate quantization and transform coding is included in our study, in which a noisy observation of source data is available at the encoder, but we are interested in estimating the unseen data at the decoder, with the help of side information.We implement a transform-domain Wyner-Ziv video coder that encodes frames independently but decodes them conditionally. Experimental results show that using the discrete cosine transform results in a rate-distortion improvement with respect to the pixel-domain coder. Transform coders of noisy images for different communication constraints are compared. Experimental results show that the noisy Wyner-Ziv transform coder achieves a performance close to the case in which the side information is also available at the encoder.

Compression of computer graphics data such as static and dynamic 3D meshes has received significant attention in recent years, since new applications require transmission over channels and storage on media with limited capacity. This includes pure graphics applications (virtual reality, games) as well as 3DTV and free viewpoint video. Efficient compression algorithms have been developed first for static 3D meshes, and later for dynamic 3D meshes and animations. Standard formats are available for instance in MPEG-4 3D mesh compression for static meshes, and Interpolator Compression for the animation part. For some important types of 3D objects, e.g. human head or body models, facial and body animation parameters have been introduced. Recent results for compression of general dynamic meshes have shown that the statistical dependencies within a mesh sequence can be exploited well by predictive coding approaches. Coders introduced so far use experimentally determined or heuristic thresholds for tuning the algorithms. In video coding, rate-distortion (RD) optimization is often used to avoid fixed thresholds and to select the optimum prediction mode. We applied these ideas and present here an RD-optimized dynamic 3D mesh coder. It includes different prediction modes as well as an RD cost computation that controls the mode selection across all possible spatial partitions of a mesh to find the clustering structure together with the associated prediction modes. The general coding structure is derived from statistical analysis of mesh sequences and exploits temporal as well as spatial mesh dependencies. To evaluate the coding efficiency of the developed coder, comparative coding results for mesh sequences at different resolutions were carried out.

We consider an interactive browsing environment for 3D scenes, which allows for the dynamic optimization of selected client views by distributing available transmission resources between geometry and texture components. Texture information is available at a server in the form of scalably compressed images, corresponding to a multitude of original image views. Surface geometry is also available at the server in the form of scalably compressed depth maps, again corresponding to a multitude of original views. Texture and depth components are both open to augmentation as more content becomes available. At any point in the interactive browsing experience, the server must decide how to allocate transmission resources between the delivery of new elements from the various original view bit-streams and new elements from the original geometry bit-streams. The proposed framework implicitly supports dynamic view sub-sampling, based on rate-distortion criteria, since the best server policy is not always to send the nearest original view image to the one which the client is rendering. In this paper, we particularly elaborate upon a novel strategy for distortion-sensitive synthesis of both geometry and rendered imagery at the client, based upon whatever data is provided by the server. We also outline how the JPIP standard for interactive communication of JPEG2000 images, can be leveraged for the 3D scene browsing application.

A theoretical framework to analyze the rate-distortion performance of a light field coding and streaming system is proposed. This framework takes into account the statistical properties of the light field images, the accuracy of the geometry information used in disparity compensation, and the prediction dependency structure or transform used to exploit correlation among views. Using this framework, the effect that various parameters have on compression efficiency is studied. The framework reveals that the efficiency gains from more accurate geometry, increase as correlation between images increases. The coding gains due to prediction suggested by the framework match those observed from experimental results. This framework is also used to study the performance of light field streaming by deriving a view-trajectory-dependent rate-distortion function. Simulation results show that the streaming results depend both the prediction structure and the viewing trajectory. For instance, independent coding of images gives the best streaming performance for certain view trajectories. These and other trends described by the simulation results agree qualitatively with actual experimental streaming results.

This paper proposes a robust video transmission algorithm, which is based on the multiple description coding (MDC) scheme that optimally splits DCT coefficients, and also proposes a switching algorithm for the optimal operation of the encoder between the single description (SD) and multiple description (MD) mode under the given packet loss rate (PLR). In the proposed MDC algorithm, DCT coefficients are either duplicated to every stream or alternated to one of streams, which is decided by the rate-distortion (RD) optimization process. The optimization problem is formulated by modifying the DCT coefficient thresholding algorithm, which can be solved by dynamic programming. Although the MD coding is more resilient than the SD coding in the severe PLR condition, its excessive redundancy results in degraded RD performance in the low PLR. To increase the scalability of MDC, we propose the SD/MD switching method, which uses the SDC for the coding efficiency in the low PLR, and MDC for the error resilience in the high PLR. An RD optimization based strategy is proposed for optimal switching under the given channel condition. Since the decoder distortion is necessary to estimate RD cost function at the time of encoding, we adopt the recursive optimal per-pixel estimate (ROPE) technique. Experimental results show that the optimal combination of the optimal split of DCT coefficients and the SD/MD switching provides higher PSNR than the conventional error resilient methods based on the inter/intra switching in low PLR conditions as well as in high PLR.

A fast H.264 Intra-prediction mode selection scheme is proposed in this work. The objective is to reduce the encoder complexity without significant rate–distortion performance degradation. The proposed method uses spatial and transform domain features of the target block jointly to filter out the majority of candidate modes. This is justified by examining the posterior error probability and the average rate–distortion loss. For the final mode selection, either the feature-based or the RDO (rate–distortion optimization)-based method is applied to 2–3 candidate modes. It is demonstrated by experimental results that the proposed scheme demands only 7–10% of the complexity of the RDO (rate–distortion optimized) mode decision scheme with little quality degradation.

In this paper, we review the rate distortion tradeoff issues in real-time video coding and introduce a minimum variation (MINVAR) distortion criterion based approach. The MINVAR based rate distortion tradeoff framework provides a local optimization strategy as a rate control mechanism in real-time video coding applications by minimizing the distortion variation while the corresponding bit rate fluctuation is limited by utilizing the encoder buffer. The proposed approach aims to achieve a smooth decoded picture quality for pleasing human visual experience. The performance of the proposed method is evaluated with H.264. The experimental results demonstrate that using the proposed approach, the decoded picture quality is smoother than the traditional H.264 joint model (JM) rate control without sacrificing global quality such that a better subjective visual quality is guaranteed.

Most constant bit-rate control algorithms aim to produce a bitstream that meets a certain bit-rate with the highest quality. Due to the non-stationary nature of the video sequence, the quality of the compressed sequence changes over time which is not desirable to end-users. In this paper, two constant-quality CBR control algorithms are proposed. They aim to produce CBR bitstream with the highest quality while having low variation in quality. Rather than controlling the quality indirectly as done in previously reported constant-quality control algorithms, they control the quality directly by using quality-matching algorithms. The frame-level Laplacian Constant Quality (FLCQ) algorithm allows one Quantization Parameter (QP) per frame, and uses a new rate-distortion model based on modeling DCT coefficients as having two-sided Laplacian distributions. The macroblock-level Viterbi Constant Quality (MVCQ) algorithm permits the QP to be changed for each macroblock using the Viterbi algorithm which reduces the search complexity. For both algorithms, besides the target bitrate, an extra degree of freedom is introduced that allows trading the variation in quality with the accuracy to the target bitrate. Simulation results show that the proposed algorithms outperform Q2 and TM5 by offering similar or higher PSNR while having lower PSNR variance.

In this paper, an efficient algorithm is proposed to reduce the computational complexity of variable-size block-matching motion estimation. We first investigate features of multiple candidate search centers, adaptive initial-blocksizes, search patterns, and search step-sizes, to match different motion characteristics and block-sizes. To avoid being trapped in local minima, the proposed algorithm uses multiple candidate motion vectors, which are obtained from different block-sizes. To further reduce the computation cost, a threshold-based early stop strategy according to the quantization parameter is suggested. With adaptive initial block-sizes, a merge-or-skip strategy is also proposed to reduce the computation for the final block-size decision. For the H.264/AVC encoder, simulations show that the proposed algorithms can speed up about 2.6–3.9 times of the original JM v6.1d encoder, which uses fast full-search for all block-sizes, and still maintain a comparable rate-distortion performance.

In the literature, several rate control techniques have been proposed to aim at the optimal quality of digitally encoded video under given bit budget, channel rate and buffer size constraints. Typically, these approaches are group-of-picture (GOP) based. For longer, heterogeneous sequences, they become unacceptably complex or struggle with model mismatches. In this paper, an off-line segment-based rate control approach is proposed for controlling the distortion variation across successive shots of a video sequence when encoding with single-layer (MPEG-4 baseline, MPEG-4 AVC) and scalable (wavelet) video codecs. Consistent quality is achieved by optimally distributing the available bits among the different segments, based on efficient rate-distortion (R-D) modelling of each segment. The individual segments are defined based on shot segmentation and activity analysis techniques. The algorithm is formulated for three different distribution models: download, progressive download and streaming. The results indicate that the proposed technique improves the quality consistency significantly, while the processing overhead compared to classical two-pass variable bit-rate (VBR) encoding is limited.

Clustering is a well-known technique for the analysis of Functional Magnetic Resonance Imaging (fMRI) data, whose main advantage is certainly flexibility: given a metric on the dataset, it "summarizes" the main characteristics of the data. But intrinsic to this approach are also the problem of defining correctly the quantization accuracy, and the number of clusters necessary to describe the data. The Information Bottleneck (IB) approach to vector quantization, proposed by Bialek and Tishby, addresses these difficulties: (1) it deals with an explicit trade-off between quantization and data fidelity; (2) it does so during the clustering procedure and not post hoc; (3) it takes into account the full statistical distribution of the features within the feature space and not only their most likely value; last, it is principled through an information theoretic formulation, which is relevant in many situations. In this paper, we present how to benefit from this method to analyze fMRI data. Our application is the clustering of voxels according to the magnitude of their responses to several experimental conditions. The IB quantization provides a consistent representation of the data, allowing for an easy interpretation and comparison of datasets.

The paper models relations of rate–ρ and QP–ρ, where ρ is defined as the percentage of zero coefficients in a frame. These two models are used for a frame-level rate-control of H.264. A linear approximation scheme is adopted to model the rate–ρ relations for the rate-control. The proposed frame-level rate-control method exploits rate-distortion optimization (RDO) to estimate macroblock modes and bitrates for the initial quantization parameter (QP). An intra-rate model is also designed to determine an initial QP for the intra-frame. In experimental results, peak signal-to-noise ratio, bitrate estimation error, rate-control accuracy in the scene changes, and computational complexity of the proposed method are analyzed for various video sequences. According to the experimental analysis, the proposed method outperforms the existing rate-control method (Joint Video Team of ISO/IEC and ITV-T the Fifth Meeting, JVT-DO69, Geneva, Switzerland, 9–17 October 2002).

This paper proposes an unequal error protection (UEP) method for MPEG-2 video transmission. Since the source and channel coders are normally concatenated, if the channel is noisy, more bits are allocated to channel coding and fewer to source coding. The situation is reversed when the channel conditions are more benign. Most of the joint source channel coding (JSCC) methods assume that the video source is subband coded, the bit error sensitivity of the source code can be modeled, and the bit allocations for different subband channels will be calculated. The UEP applied to different subbands is the rate compatible punctured convolution channel coder. However, the MPEG-2 coding is not a subband coding, the bit error sensitivity function for the coded video can no longer be applied. Here, we develop a different method to find the rate-distortion functions for JSCC of the MPEG-2 video. In the experiments, we show that the end-to-end distortion of our UEP method is smaller than the equal error protection method for the same total bit-rate.

Many encoder architectures, notably for scalable systems such as MPEG-2 SNR scalability and stream morphing, require DCT-domain techniques for the estimation of the rate-distortion characteristics of a video frame for use in the rate control process. This paper develops a CBR rate control framework that can be used for single-layer MPEG-4 Simple Profile as well as two MPEG-4-based scalable systems, one following the structure described in MPEG-2 and also stream morphing. The framework supports perceptual quantization and also low bit rate operation by modelling macroblock overhead as well as the bits required to code non-zero DCT coefficients. Experimental results demonstrate the performance of stream morphing, in both objective and subjective measures, is in almost all cases superior to the MPEG-2 SNR scalability and MPEG-4 Fine Granularity Scalability.

This paper proposes several enhancements to the Set Partitioning in Hierarchical Trees (SPIHT) image coding algorithm without changing the original algorithm's general skeleton. First and foremost, a method for significance map pruning based on a rate-distortion criterion is introduced. Specifically, the (Type A) sets of wavelet coefficients with small ratios of estimated distortion reduction to estimated rate contribution are deemed insignificant and effectively pruned. Even though determining such sets requires the computational complexity of the encoder to increase considerably with respect to the original SPIHT encoder, the original SPIHT decoder may still be used to decode the generated bitstream with a low computational complexity. The paper also proposes three low complexity enhancements by more sophisticated use of the adaptive arithmetic coder. Simulation results demonstrate that all these enhancements yield modest compression gains at moderate to high rates. Publisher's Version Q2 WOS:000186655800001

In many video based applications, it is essential to precisely control the bit rate of video streams for transmission over different networks and channel bandwidth. One critical element in a bit rate control algorithm is the bit production model that predicts the number of produced bits when a certain quantization parameter is used. In this paper, we present a novel bit allocation and rate control algorithm for compressed domain video transcoding. The specific transcoding issue mentioned is referred to as bit rate adaptation or rate shaping. We first review the architectures of different bit rate adaptation transcoders and the generic rate control problem. Then, we propose and formulate an approximate linear bit allocation model, which is based on experimental results. Based on this model, we propose an adaptive bit estimation and allocation scheme for video transcoding. Implementation results show that the proposed algorithm can provide accurate bit allocation, stable buffer occupancy and improved video quality as compared to existing approaches. This rate control scheme can be used to provide flexible video bit rate adaptation and stable transmission of video streams over heterogeneous networks.