Cooperative Caching for Scalable Video Coding Using Value-Decomposed Dimensional Networks

Scalable video coding(SVC)has been widely used in video-on-demand(VOD)service,to efficiently satisfy users’different video quality requirements and dynamically adjust video stream to timevariant wireless channels.Under the 5G network structure,we consider a cooperative caching scheme inside each cluster with SVC to economically utilize the limited caching storage.A novel multi-agent deep reinforcement learning(MADRL)framework is proposed to jointly optimize the video access delay and users’satisfaction,where an aggregation node is introduced helping individual agents to achieve global observations and overall system rewards.Moreover,to cope with the large action space caused by the large number of videos and users,a dimension decomposition method is embedded into the neural network in each agent,which greatly reduce the computational complexity and memory cost of the reinforcement learning.Experimental results show that:1)the proposed value-decomposed dimensional network(VDDN)algorithm achieves an obvious performance gain versus the traditional MADRL;2)the proposed VDDN algorithm can handle an extremely large action space and quickly converge with a low computational complexity.

Fast mode decision for Hierarchical-B pictures in scalable video coding

To decrease the computational complexity of adaptive inter-layer prediction and improve the encoding efficiency in sealable video coding, a mode decision algorithm is proposed by exploiting the part of used candidate modes of the co-located reference macrobloeks for Hierarchical-B pictures. This scheme reduces the amount of the candidate modes to generate a dynamic list for the current encoding macroblock according to the statistical information derived from the co-located reference macroblocks in different temporal levels. The experimental results show that this fast algorithm reduces approximately 31% encoding time on average with the negligible loss of encoding performance.

Trade-off Optimization for Scalable Video Coding Streaming in Relay-Based OFDMA Networks

To achieve an optimal trade-off between video quality and energy efficiency in the uplink streaming of multi-user Scalable Video Coding (SVC) videos in relay-based Orthogonal Frequency Division Multiple Access (OFDMA) cellular networks, a cross-layer design framework that jointly selects the Transmission Policy (TP) for SVC video frames, assigns OFDMA subcarriers, and allocates power for each subcarrier is proposed. We apply the dual decomposition method to the problem, and obtain a TP selection subproblem for each SVC video adaptation and a resource allocation subproblem of Joint Subcarrier, Relay and Power Allocation (JSRPA). A second level of dual decomposition is used to divide the JSRPA problem into independent subcarrier subproblems. The proposed Crosslayer Trade-off Optimization (CTO) algorithm is sub-distributed with significantly low complexity. A performance evaluation with typical SVC video traces demonstrates that the proposed algorithm is able to converge and efficiently achieve the optimal trade-off between the video quality and energy consumption at the MSs for uplink SVC streaming.

Efficient Multiple Description Scalable Video Coding Scheme Based on Weighted Signal Combinations

A simple, efficient multiple description coding （MDC） algorithm was developed based on weighted signal combinations. The scheme uses the standard video encoder with a pre-processing stage to generate multiple descriptions. The decoder then uses a post-processing algorithm to combine the descriptions to provide better image quality. A scalable codec in the MDC system allows the system to provide multiple descriptions and scalability at the same time. In addition, since the different scalable descriptions may have different qualities, a simple averaging process is not optimal. An optimal weighted combination of the two descriptions was developed based on the signal to noise ratios. Compared with the simple average combination, the algorithm significantly improved the video quality, especially with large quality differences between the two descriptions, with gains of up to 3.56 dB.

Component-adaptive up-sampling for inter layer interpolation in scalable video coding

Scalable video coding （SVC） is a newly emerging standard to be finalized as an extension of H.264/AVC. The most attractive characters in SVC are the inter layer prediction techniques, such as Intra_BL mode. But in current SVC scheme, a uniform up-sampling filter （UUSF） is employed to magnify all components of an image, which will be very inefficient and result in a lot of redundant computational complexity. To overcome this, we propose an efficient component-adaptive up-sampling filter （CAUSF） for inter layer interpolation. In CAUSF, one character of human vision system is considered, and different up- sampling filters are assigned to different components. In particular, the six-tap FIR filter used in UUSF is kept and assigned for luminance component. But for chrominance components, a new four-tap FIR filter is used. Experimental results show that CAUSF maintains the performances of coded bit-rate and PSNR-Y without any noticeable loss, and provides significant reduction in computational complexity.

Flexible three-band motion-compensated temporal filtering for scalable video coding

A novel scheme for scalable video coding using three-band lifting-based motion-compensated transform is presented in this article. A series of flexible three-band motion-compensated lifting steps are used to implement the temporal wavelet transform, which provide improved compression performance by selecting specific motion model according to real video sequences, and offer higher temporal scalability flexibility by using three-band lifting steps. The experimental results compared with motion picture expert group （MPEG）-4 codec concerning standard video sequences demonstrate the effectiveness of the method.

A pixel-based outlier-free motion estimation algorithm for scalable video quality enhancement

Scalable video quality enhancement refers to the process of enhancing low quality frames using high quality ones in scalable video bitstreams with time-varying qualities. A key problem in the enhancement is how to search for correspondence between high quality and low quality frames. Previous algorithms usually use block-based motion estimation to search for correspondences. Such an approach can hardly estimate scale and rotation transforms and always in- troduces outliers to the motion estimation results. In this paper, we propose a pixel-based outlier-free motion estimation algorithm to solve this problem. In our algorithm, the motion vector for each pixel is calculated with respect to estimate translation, scale, and rotation transforms. The motion relationships between neighboring pixels are considered via the Markov random field model to improve the motion estimation accuracy. Outliers are detected and avoided by taking both blocking effects and matching percentage in scale- invariant feature transform field into consideration. Experiments are conducted in two scenarios that exhibit spatial scalability and quality scalability, respectively. Experimental results demonstrate that, in comparison with previous algorithms, the proposed algorithm achieves better correspondence and avoids the simultaneous introduction of outliers, especially for videos with scale and rotation transforms.

Genetic-optimization framework for SVC transmission based on partial cooperative communication

A genetic-optimization framework based on the partial cooperation communication protocol is proposed for scalable video coding (SVC) stream transmission under multi-relay amplify and forward cooperative networks. Unlike traditional cooperative transmission schemes, the transmission mode for each coded unit in this new protocol can be switched flexibly between direct transmission and cooperative transmission. Obviously, under this protocol, the bandwidth efficiency and transmission robustness can be balanced adaptively according to the priority level of coded units and wireless channel fading characteristics. Based on this, a well-known genetic optimization algorithm-differential evolution is exploited here to find the jointly optimal transmission modes, power allocation and unequal error protection (UEP) channel coding strategies to minimize the end to end reconstructed video distortion. Extensive simulation results show that, compared with classical optimal cooperative UEP transmission schemes, the proposed optimized transmission framework based on the partial cooperative protocol can bring significant peak-signal-to-noise-ratio (PSNR) gains for the reconstructed video in a variety of channel bandwidth, power budget and test sequences.

A video conferencing system based on SDN-enabled SVC multicast

Current typical video conferencing connection is bridged by a multipoint control unit(MCU),which may cause large delay and communication bottleneck for the whole system.With the development of network technology,a video conferencing system can be implemented based on software-defined networking(SDN),which makes the service controllable and improves the scalability and flexibility.Additionally,a video encoding method called scalable video coding(SVC) can also help.In this paper,we propose a video conferencing architecture based on SDN-enabled SVC multicasting,which discards the traditional Internet group management protocol(IGMP) and MCU.The system implements SVC multicast streaming to satisfy different device capabilities of various conference terminals.The SDN controller is responsible for dynamically managing and controlling the layers of a video stream when a conference member faces network congestion.Also,a conference manager is designed to facilitate the management of the conference members.Experimental results show that our system can not only provide a flexible and controllable video delivery,but also reduce the network usage while guaranteeing the quality of service(QoS) of video conferencing.

H.264/SVC Mode Decision Based on Mode Correlation and Desired Mode List

Design of video encoders involves implementation of fast mode decision(FMD) algorithm to reduce computation complexity while maintaining the performance of the coding. Although H.264/scalable video coding(SVC) achieves high scalability and coding efficiency, it also has high complexity in implementing its exhaustive computation. In this paper, a novel algorithm is proposed to reduce the redundant candidate modes by making use of the correlation among layers. A desired mode list is created based on the probability to be the best mode for each block in base layer and a candidate mode selection in the enhancement layer by the correlations of modes among reference frame and current frame. Our algorithm is implemented in joint scalable video model(JSVM)9.19.15 reference software and the performance is evaluated based on the average encoding time, peak signal to noise ration(PSNR)and bit rate. The experimental results show 41.89% improvement in encoding time with minimal loss of 0.02 dB in PSNR and 0.05%increase in bit rate.

A Stochastic Optimization Approach for Dynamic Adaptive Streaming over NDN

Nowadays,video streaming counts for the major part of network traffic over the Internet.However,on account of the host-to-host mechanism of the traditional IP network,video distribution over IP-based Internet encounters bottlenecks.Fortunately,a new proposed future Internet architecture,named data networking(NDN)can improve the performance of video distribution by its features such as in-network storage,multi-path forwarding,etc.In this paper,we design an adaptive bitrate algorithm based on Lyapunov optimization theory over NDN to optimize the long-term quality-of-experience(QoE)of video distribution while ensuring the stability of the whole system.When the network condition is abundant and stable,the problem can be simplified by approximating to a fixed-slot queuing model,but the theoretical performance will degrade when the network status is poor and fluctuate fiercely.Therefore,we divide the problem into two models of fixed time slot and non-fixed time slot and design two Lyapunov optimization algorithms to adapt different network scenarios.The proposed algorithms do not require prior knowledge of the network bandwidth and are capable of running online with the client’s available information.Simulation and realistic experiment results demonstrate that our algorithms perform better than others in NDN.