Video compression

Introduction

Video coding and processing group focuses its research on multiple aspects of 2D and 3D video coding for transmission over a variety of communication networks. Among the most significant areas of the expertise are novel compression schemes applied to videos that are based on widely deployed international video coding standards, such as H.264/AVC, Multi-View Coding (MVC) and more recently High Efficiency Video Coding (HEVC) and its extensions including Multiview (MV-HECV), 3 Dimension (3D-HEVC), Scalability (SHVC).

Our contributions include the advancement in the error robustness schemes of existing encoders, making use of the statistical dependencies between different elements of 3D media, video layers, and improvements in the general rate-distortion efficiency for targeted applications.

We also target our fundamental research on future video coding techniques like Distributed Video Coding (DVC) as well as the development of video transmission systems and video quality assessment for oriental video applications.

People

  • Dr. Dinh Trieu Duong from University of Engineering and Technology, VNUH. He got Ph.D degree in Korea University, Korea. His research includes Telecommunication, Video Coding and Communication.
  • Dr. Nguyen Huu Tien from Posts and Telecommunications Institute of Technology. He got Ph.D degree in Chulalongkorn University, Thailand. His research includes Image Processing and Video Coding.
  • Dr. Hoang Van Xiem from University of Engineering and Technology, VNUH. He got Ph.D in Instituto de Telecomunicações, Portugal. His research includes Video Coding and Image Processing.
  • Assoc. Prof. Dr. Le Thanh Ha from University of Engineering and Technology, VNUH. He got Ph.D degree in Korea University, Korea. His research includes Video Coding and Communication, Image Processing, Computer Vision and Satellite Image Processing.

Ongoing projects

  • Video analysis and processing methods for surveillance cameras, 2016-2017, VP9 Company.
  • Research and develop next generation distributed scalable video coding scheme for multimedia applications, 2017 – 2019, Nafosted.
  • Developing a Distributed video compressor/De-compressor DVC, and Optimizing the Performance of DVC Deployed in Wireless Visual Sensor Networks.
  • Coding and communication of multiview video plus depth for 3D Television Systems, 2013-2016, Nafosted.

Publication

  • Xiem Hoang Van, João Ascenso, and Fernando Pereira, Adaptive Scalable Video Coding: a HEVC based Framework Combining the Predictive and Distributed Paradigms, IEEE Transaction on Circuits and Systems for Video Technology, IF: 2.65 (SCI), Mar. 2016.
  • Xiem Hoang Van, Dinh Trieu Duong, and Le Thanh Ha, Spatial-Temporal Feature Extraction based Adaptive Search Range for Effective Frame Rate-Up Conversion, IEEE Advanced Technologies for Communication, Hanoi, Vietnam, Oct. 2016.
  • Minh Le Dinh, Long Vuong Tung, Xiem Hoang Van, Dinh Trieu Duong, Tung Pham Thanh, and Ha Le Thanh, Improving 3DTV View Synthesis using Motion Compensated Temporal Interpolation, IEEE Advanced Technologies for Communication, Hanoi, Vietnam, Oct. 2016.
  • Thao Nguyen Thi Huong, Huu-Tien Vu, Xiem Hoang Van, Ha Le Thanh, and Dinh Trieu Duong, Side Information Creation using Adaptive Block size for Distributed Video Coding, IEEE Advanced Technologies for Communication, Hanoi, Vietnam, Oct. 2016.
  • Long Vuong Tung, Minh Le Dinh, Xiem HoangVan, Dinh Trieu Duong, Tien Huu Vu, and Ha Le Thanh, View Synthesis Method for 3D Television based on Temporal and Disparity Correlation, EURASIP Journal on Image and Video processing (Submitted).
  • Thanh Ha Le, Vuong Tung Long, Dinh Trieu Duong, and Seung-Won Jung, Reduced Reference Quality Metric for Synthesized Virtual Views in 3DTV, ETRI Journal vol. 38, no. 6, pp. 1114-1123, Dec., 2016.
  • Thanh Ha Le, Seung-Won Jung, Chee Sun Won, A New Depth Image Quality Metric Using a Pair of Color and Depth Images, Multimedia Tools and Applications, 2016.
  • Pham Thanh Nam, Vu Duy Khuong, Dinh Trieu Duong, and Thanh Ha Le, Efficient Region-of-Interest Based Adaptive Bit Allocation for 3D-TV Video Transmission over Networks, VNU Journal of Computer Science and Communication Engineering, Vol. 32, No.1, pp. 1-9, 2016.

Some selected research topics and results

New Generation Distributed Scalable Video Coding framework

Since H.265/HEVC is the most recent video coding standard, to be replaced for the widely used H.264/AVC standard, new generation scalable video coding model will be a scalable extension of the H.265/HEVC. Similar to traditional scalable video coding models, video information is encoded and decoded following a layered coding structure including a base and one or more enhancement layers. One of major features of new generation scalable video coding is that the base layer must be compatible with not only the prior H.264/AVC but also the H.265/HEVC standards as illustrated in Figure below. In this case, the new generation DSVC is able to decode the video encoded using H.265/HEVC.

shvc

The current research achievements of scalable video coding standards are quite limited, notably in emerging multimedia applications such as visual sensor networks, video surveillance systems, and remote space transmission, which usually require low energy consumption of video coding for long time usage and the error prone adaptation for seamlessly transmitting and receiving video. To solve this problem, we propose in this research a novel new generation distributed scalable video coding (NG-DSVC) framework, which combines the most recent H.265/HEVC standard with the promising distributed video coding (DVC) model. The proposed NG-DSVC framework aims to utilize the coding effectiveness of both H.265/HEVC and DVC for the satisfaction of low energy consumption and high error tolerance.

Adaptive FMO based error resilience in H.264 video coding

In this work, we proposed an error resilience scheme for wireless video coding based on adaptive Flexible Macroblock Ordering (FMO) and intra refresh.  FMO explicit map is generated frame-by-frame using the estimated prior information.  The prior information involves the estimation of locations of guards and bursts of errors in the channel using three-state Markov model and the estimation of the error propagation effects (EEP) from the previous frame to the current frame. In addition, the role of the current frame in propagating error to the next frame is also considered. Intra refresh with a suitable rate which is adaptive to channel state is used to reduce the dependence between frames and thus the effect of error propagation is stopped.

fmo

Experimental results show that our proposed method gains some improvements in terms of PSNR as compared to some other methods that have not taken channel condition and error propagation into consideration in generating FMO map.

Reduced reference quality metric for depth image

Typical depth quality metrics require the ground truth depth image or stereoscopic color image pair, which are not always available in many practical applications. In this work, we propose a new depth image quality metric which demands only a single pair of color and depth images.

rr-dqm

Our observations reveal that the depth distortion is strongly related to the local image characteristics, which in turn leads us to formulate a new distortion assessment method for the edge and non-edge pixels in the depth image. The local depth distortion is adaptively weighted using the Gabor filtered color image and added up to the global depth image quality metric. The experimental results show that the proposed metric closely approximates the depth quality metrics that use the ground truth depth or stereo color image pair.

Reduced reference quality metric for synthesized virtual views in 3DTV

In 3DTV, multiple-view video plus depth (MVD) has been widely used due to its effectiveness in 3D data representation. Based on the MVD method, color videos with only a limited number of real viewpoints are compressed and transmitted along with captured or estimated depth videos. Since synthesized views are generated from decoded real views, their original reference views do not exist either at the transmitter or receiver. Therefore, it is challenging to define an efficient metric to evaluate the quality of synthesized images.

rr-vvqm

We propose a novel metric named reduced-reference quality metric for synthesized views (RR-VVQM). First, the effects of depth distortion on the quality of synthesized images are analyzed and experimented in detail. Then, we employ the high correlation between local depth distortions and local color characteristics of the decoded depth and color images, respectively, to achieve an efficient depth quality metric for each real view. Finally, the objective quality metric of synthesized views is obtained by combining all the depth quality metrics obtained from the decoded real views. The experimental results show that the proposed quality metric has excellent correlations with full reference image and video quality metrics.

Demo

Coming soon

Software

Coming soon