Deep chroma prediction of Wyner–Ziv frames in distributed video coding of wireless capsule endoscopy video

Compression of captured video frames is crucial for saving the power in wireless capsule endoscopy (WCE). A low complexity encoder is desired to limit the power consumption required for compressing the WCE video. Distributed video coding (DVC) technique is best suitable for designing a low complexity encoder. In this technique, frames captured in RGB colour space are converted into YCbCr colour space. Both Y and CbCr representing luma and chroma components of the Wyner–Ziv (WZ) frames are processed and encoded in existing DVC techniques proposed for WCE video compression. In the WCE video, consecutive frames exhibit more similarity in texture and colour properties. The proposed work uses these properties to present a method for processing and encoding only the luma component of a WZ frame. The chroma components of the WZ frame are predicted by an encoder–decoder based deep chroma prediction model at the decoder by matching luma and texture information of the keyframe and WZ frame. The proposed method reduces the computations required for encoding and transmitting of WZ chroma component. The results show that the proposed DVC with a deep chroma prediction model performs better when compared to motion JPEG and existing DVC systems for WCE at the reduced encoder complexity.     

Perceptual-based distributed video coding

In this paper, we propose a perceptual-based distributed video coding (DVC) technique. Unlike traditional video codecs, DVC applies video prediction process at the decoder side using previously received frames. The predicted video frames (i.e., side information) contain prediction errors. The encoder then transmits error-correcting parity bits to the decoder to reconstruct the video frames from side information. However, channel codes based on i.i.d. noise models are not always efficient in correcting video prediction errors. In addition, some of the prediction errors do not cause perceptible visual distortions. From perceptual coding point of view, there is no need to correct such errors. This paper proposes a scheme for the decoder to perform perceptual quality analysis on the predicted side information. The decoder only requests parity bits to correct visually sensitive errors. More importantly, with the proposed technique, key frames can be encoded at higher rates while still maintaining consistent visual quality across the video sequence. As a result, even the objective PSNR measure of the decoded video sequence will increase too. Experimental results show that the proposed technique improves the R-D performance of a transform domain DVC codec both subjectively and objectively. Comparisons with a well-known DVC codec show that the proposed perceptual-based DVC coding scheme is very promising for distributed video coding framework.     

一种空间域Wyner-Ziv视频编码系统的性能改进算法

分布式视频编码是建立在Slepian-Wolf和Wyner-Ziv信息编码理论基础上的全新视频编码框架,具有编码复杂度低,编码效率较高,抗误码性能好的特点.本文首先简单介绍了一种典型的分布式视频编码实现方案——空间域Wyner-Ziv视频编码,随后提出一种空间域Wyner-Ziv视频编码系统的性能改进算法,该算法在不增加编码复杂度的基础上,在解码端利用双向运动估计预测获取更高质量的边信息,同时采用基于Huber-Markov随机场约束的联合迭代解码算法重建图像.实验结果表明,在相同的输出码流情况下,本文改进算法在解码端重建图像的峰值信噪比与空间域Wyner-Ziv视频编码算法相比平均提高2dB,并且主观效果有所改善.     

Advanced side information creation techniques and framework for Wyner–Ziv video coding

Recently, several distributed video coding (DVC) solutions based on the distributed source coding (DSC) paradigm have appeared in the literature. Wyner–Ziv (WZ) video coding, a particular case of DVC where side information is made available at the decoder, enable to achieve a flexible distribution of the computational complexity between the encoder and decoder, promising to fulfill novel requirements from applications such as video surveillance, sensor networks and mobile camera phones. The quality of the side information at the decoder has a critical role in determining the WZ video coding rate-distortion (RD) performance, notably to raise it to a level as close as possible to the RD performance of standard predictive video coding schemes. Towards this target, efficient motion search algorithms for powerful frame interpolation are much needed at the decoder. In this paper, the RD performance of a Wyner–Ziv video codec is improved by using novel, advanced motion compensated frame interpolation techniques to generate the side information. The development of these type of side information estimators is a difficult problem in WZ video coding, especially because the decoder only has available some reference, decoded frames. Based on the regularization of the motion field, novel side information creation techniques are proposed in this paper along with a new frame interpolation framework able to generate higher quality side information at the decoder. To illustrate the RD performance improvements, this novel side information creation framework has been integrated in a transform domain turbo coding based Wyner–Ziv video codec. Experimental results show that the novel side information creation solution leads to better RD performance than available state-of-the-art side information estimators, with improvements up to 2 dB; moreover, it allows outperforming H.264/AVC Intra by up to 3 dB with a lower encoding complexity.     

Independent key frame coding using correlated pixels in distributed video coding

A novel intra-coding technique is proposed that eliminates the requirement of a secondary coding scheme for coding the key frames in distributed video coding (DVC). The proposed technique uses the Slepian-Wolf theorem and Wyner-Ziv (WZ) coding with spatially predicted information to transmit the key-frames to the DVC decoder. Simulation results show that the proposed WZ-intra coding technique (WZ-I) can achieve up to 5 dB PSNR gain compared to MPEG-2 intra coding (MPEG-I) at the same bit rate with negligible computational cost to the encoder     

Turbo trellis coded modulation for transform domain distributed video coding

A transform domain distributed video coding (DVC) codec is proposed using turbo trellis coded modulation (TTCM). TTCM symbols are generated at the DVC decoder using the side information and the parity bits received from the DVC encoder. These generated symbols are used at the TTCM-based DVC decoder to decode the bit stream. Simulation results show that a significant rate-distortion performance gain can be achieved using the proposed codec compared to the best state-of-the-art transform domain DVC codecs discussed in the literature.     

Modified VQ-BTC algorithm for image compression

VQ-BTC is a recent technique used in the coding of image data to combat edge degradation produced by vector quantisation (VQ) or block truncation coding (BTC). However, it has high encoding complexity and needs a large amount of memory to store 31 codebooks at both the encoder and decoder. A modified VQ-BTC (MVQ-BTC) algorithm is presented which achieves a performance close to that of VQ-BTC, but needs only three codebooks, and requires less computation time than VQ-BTC     

Flexible distribution of complexity by hybrid predictive-distributed video coding

There is currently limited flexibility for distributing complexity in a video coding system. While rate-distortion-complexity (RDC) optimization techniques have been proposed for conventional predictive video coding with encoder-side motion estimation, they fail to offer true flexible distribution of complexity between encoder and decoder since the encoder is assumed to have always more computational resources available than the decoder. On the other hand, distributed video coding solutions with decoder-side motion estimation have been proposed, but hardly any RDC optimized systems have been developed.To offer more flexibility for video applications involving multi-tasking or battery-constrained devices, in this paper, we propose a codec combining predictive video coding concepts and techniques from distributed video coding and show the flexibility of this method in distributing complexity. We propose several modes to code frames, and provide complexity analysis illustrating encoder and decoder computational complexity for each mode. Rate distortion results for each mode indicate that the coding efficiency is similar. We describe a method to choose which mode to use for coding each inter frame, taking into account encoder and decoder complexity constraints, and illustrate how complexity is distributed more flexibly.     

Multiple side information streams for distributed video coding

An improved Wyner-Ziv decoder for distributed video coding (DVC) is proposed, which uses multiple side information streams obtained by using multiple reference frames. Simulation results show that the proposed algorithm can achieve a significant PSNR gain of up to 2.4 dB over the best available DVC codec at the same bit rate     

基于边信息的分布式视频压缩感知的残差重构

压缩感知(Compressed Sensing,CS)结合了视频信号的变换和信息压缩过程,为简化编码算法提供了一种新的解决思路.把分布式视频编码(DVC)和CS结合在一起,构建简单的视频编码框架,并利用原始视频帧与边信息之间的相关性进行残差重构,提出了一种基于边信息的分布式视频压缩感知编解码方案.此方法对关键帧采用传统的帧内编、解码;对非关键帧CS进行随机观测提取观测向量,解码端利用优化的边信息和传输的CS观测向量进行联合重构.实验结果表明,该方法在运动较平滑的序列中比参考方案的恢复质量提高了4 ～6 dB.     

Low delay distributed video coding with refined side information

Distributed video coding (DVC) is a new video coding paradigm based upon two fundamental theoretical results: the Slepian–Wolf and Wyner–Ziv theorems. Among other benefits, this new coding paradigm may allow a flexible complexity allocation between the encoder and the decoder. Several DVC codecs have been developed over the years addressing the specific requirements of emerging applications such as wireless video surveillance and sensor networks. While state-of-the-art DVC codecs, such as the DISCOVER DVC codec, have shown promising RD performance, most DVC codecs in the literature do not consider low delay requirements which are relevant for some of the addressed applications. In this context, the main objective and novelty of this paper is to propose an efficient, low delay and fully practical DVC codec based on the Stanford DVC architecture adopting a side information iterative refinement approach. The obtained performance results show that the developed DVC solution fulfils the objectives regarding relevant benchmarks, notably due to the novel side information creation and correlation noise modeling tools integrated in a side information iterative refinement framework.     

Rate-distortion driven decoder-side bitplane mode decision for distributed video coding

Distributed video coding (DVC) features simple encoders but complex decoders, which lies in contrast to conventional video compression solutions such as H.264/AVC. This shift in complexity is realized by performing motion estimation at the decoder side instead of at the encoder, which brings a number of problems that need to be dealt with. One of these problems is that, while employing different coding modes yields significant coding gains in classical video compression systems, it is still difficult to fully exploit this in DVC without increasing the complexity at the encoder side. Therefore, in this paper, instead of using an encoder-side approach, techniques for decoder-side mode decision are proposed. A rate-distortion model is derived that takes into account the position of the side information in the quantization bin. This model is then used to perform mode decision at the coefficient level and bitplane level. Average rate gains of 13–28% over the state-of-the-art DISCOVER codec are reported, for a GOP of size four, for several test sequences.     

Low complexity distributed video coding

ContextConventional video encoding is a computationally intensive process that requires a lot of computing resources, power and memory. Such codecs cannot be deployed in remote sensors that are constrained in terms of power, memory and computational capabilities. For such applications, distributed video coding might hold the answer.ObjectiveIn this paper, we propose a distributed video coding (DVC) architecture that adheres to the principles of DVC by shifting the computational complexity from the encoder to the decoder and caters to low-motion scenarios like video conferencing and surveillance of hallways and buildings.MethodThe architecture presented is block-based and introduces a simple yet effective classification scheme that aims at maximizing the use of skip blocks to exploit temporal correlation between consecutive frames. In addition to the skip blocks, a dynamic GOP size control algorithm is proposed that instantaneously alters the GOP size in response to the video statistics without causing any latency and without the need to buffer additional frames at the encoder. To facilitate real-time video delivery and consumption, iterative channel codes like low density parity check codes and turbo codes are not used and in their place a Bose–Chaudhuri–Hocquenghem (BCH) code with encoder rate control is used.ResultsIn spite of reducing the complexity and eliminating the feedback channel, the proposed architecture can match and even surpass the performance of current DVC systems making it a viable solution as a codec for low-motion scenarios.ConclusionWe conclude that the proposed architecture is a suitable solution for applications that require real-time, low bit rate video transmission but have constrained resources and cannot support the complex conventional video encoding solutions.Practical implicationsThe practical implications of the proposed DVC architecture include deployment in remote video sensors like hallway and building surveillance, video conferencing, video sensors that are deployed in remote regions (wildlife surveillance applications), and capsule endoscopy.     

Gain-Adaptive Vector Quantization with Application to Speech Coding

The generalization of gain adaptation to vector quantization (VQ) is explored in this paper and a comprehensive examination of alternative techniques is presented. We introduce a class of adaptive vector quantizers that can dynamically adjust the "gain" or amplitude scale of code vectors according to the input signal level. The encoder uses a gain estimator to determine a suitable normalization of each input vector prior to VQ encoding. The normalized vectors have reduced dynamic range and can then be more efficiently coded. At the receiver, the VQ decoder output is multiplied by the estimated gain. Both forward and backward adaptation are considered and several different gain estimators are compared and evaluated. Gain-adaptive VQ can be used alone for "vector PCM" coding (i.e., direct waveform VQ) or as a building block in other vector coding schemes. The design algorithm for generating the appropriate gain-normalized VQ codebook is introduced. When applied to speech coding, gain-adaptive VQ achieves significant performance improvement over fixed VQ with a negligible increase in complexity.     

Introducing skip mode in distributed video coding

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%.     

Side information estimation and new symmetric schemes for multi-view distributed video coding

This paper deals with distributed video coding (DVC) for multi-view sequences. DVC of multi-view sequences is a recent field of research, with huge potential impact in applications such as videosurveillance, real-time event streaming from multiple cameras, and, in general, immersive communications. It raises however several problems, and in this paper we tackle two of them. Based on the principles of Wyner–Ziv (WZ) coding, in multi-view DVC many estimations can be generated in order to create the side information (SI) at the decoder. It has been shown that the quality of the SI strongly influences the global coding performances. Therefore, this paper proposes to study the contribution of multiple SI estimations (in the temporal and view directions) to the global performances. Moreover, we propose new symmetric schemes for longer group of pictures (GOP) in multi-view DVC and show that we can further exploit the long-term correlations using a new kind of estimation, called diagonal. For such schemes, several decoding strategies may be envisaged. We perform a theoretical study of the temporal and inter-view dependencies, and confirm by experiments the conclusion about the best decoding strategy.     

Reduced decoder complexity and latency in pixel-domain Wyner–Ziv video coders

In some video coding applications, it is desirable to reduce the complexity of the video encoder at the expense of a more complex decoder. Wyner–Ziv (WZ) video coding is a new paradigm that aims to achieve this. To allocate a proper number of bits to each frame, most WZ video coding algorithms use a feedback channel, which allows the decoder to request additional bits when needed. However, due to these multiple bit requests, the complexity and the latency of WZ video decoders increase massively. To overcome these problems, in this paper we propose a rate allocation (RA) algorithm for pixel-domain WZ video coders. This algorithm estimates at the encoder the number of bits needed for the decoding of every frame while still keeping the encoder complexity low. Experimental results show that, by using our RA algorithm, the number of bit requests over the feedback channel—and hence, the decoder complexity and the latency—are significantly reduced. Meanwhile, a very near-to-optimal rate-distortion performance is maintained. This work has been partially supported by the Spanish Ministry of Education and Science and the European Commission (FEDER) under grant TEC2005-07751-C02-01. A. Pižurica is a postdoctoral research fellow of FWO, Flanders.     

Error-resilient scalable compression based on distributed video coding

Distributed Video Coding (DVC) is a new paradigm for video compression based on the information theoretical results of Slepian–Wolf (SW) and Wyner–Ziv (WZ). In this work, a performance analysis of image and video coding schemes based on DVC is presented, addressing temporal, quality and spatial scalability. More specifically, conventional coding is used to obtain a base layer while WZ coding generates the enhancement layers. At the decoder, the base layer is used to construct Side Information (SI) for the DVC decoding process. Initially, we show that the scalable DVC approach is codec-independent, which means that it is independent from the method used to encode the base layer. Moreover, the influence of the base layer quality on the overall performance of the schemes is studied. Finally, evaluation of the proposed schemes is performed in both cases, with and without transmission errors. The simulation results show that scalable DVC has a lower compression efficiency than conventional scalable coding (i.e. scalable video coding and JPEG2000 for video and image, respectively) in error-free conditions. On the other hand, the DVC-based schemes show better error resilience as they outperform conventional scalable coding in error-prone conditions. More specifically, the Rate Distortion (RD) performance of the proposed schemes for image coding is compared with respect to Reed Solomon (RS) protected JPEG2000. While the latter exhibits a cliff effect as its performance dramatically decreases after a certain error rate, the performance of the DVC-based schemes decreases in a steady way with error rate increase.     

An improved vector quantization method using deep neural network

To address the challenging problem of vector quantization (VQ) for high dimensional vector using large coding bits, this work proposes a novel deep neural network (DNN) based VQ method. This method uses a k-means based vector quantizer as an encoder and a DNN as a decoder. The decoder is initialized by the decoder network of deep auto-encoder, fed with the codes provided by the k-means based vector quantizer, and trained to minimize the coding error of VQ system. Experiments on speech spectrogram coding demonstrate that, compared with the k-means based method and a recently introduced DNN-based method, the proposed method significantly reduces the coding error. Furthermore, in the experiments of coding multi-frame speech spectrogram, the proposed method achieves about 11% relative gain over the k-means based method in terms of segmental signal to noise ratio (SegSNR).     

Improved side information generation algorithm for Wyner-Ziv video coding简

Side information has a significant influence on the rate-distortion（RD） performance of distributed video coding（DVC）. In the conventional motion compensated frame interpolation scheme, all blocks adopt the same side-information generation method regardless of the motion intensity inequality at different regions. In this paper, an improved method is proposed. The image blocks are classified into two modes, fast motion and slow motion, by simply computing the discrete cosine transformation（DCT） coefficients at the encoder. On the decoder, it chooses the direct interpolation and refined motion compensated interpolation correspondingly to generate side information. Experimental results show that the proposed method, without increasing the encoder complexity, can increase the average peak signal-to-noise ratio（PSNR） by up to 1~ 2 dB compared with the existing algorithm. Meanwhile, the proposed algorithm significantly improves the subjective quality of the side information.