Coding scheme with skip mode based on motion filed detection for DVC

Abstract

ABSTRACT Motion estimation has been shifted from encoder to the decoder in distributed video coding (DVC). In this paper, a simplified skip mode is introduced in the WZ (Wyner-Ziv) frame coding process. In the proposed scheme, a skip mode decision process is performed to determine whether to apply skip mode to the blocks with low motion first. With the skip-mode, the block is reconstructed from the side informati on in the decoder without any encoding bits. In addition, the non-skip mode blocks are divided into two parts by motion activity. The encoding bitplane is extracted within each set and encoded independently. Simulations show the proposed scheme can achieve up to 54.29% bitrate savings without visible PSNR sacrifice Keywords: Distributed video coding, Wyner-Ziv coding, skip mode 1. INTRODUCTION The traditional coding standards are challenged by the emerging applications characterized by a significantly lower complex encoder and a high complex video decoder structure. However, distributed video coding is ideally meeting this requirement. The theory of distributed source coding was proposed in 1970s by Slepian-Wolf [1] and Wyner-Ziv [2]. Compared with the traditional coding codec, a dramatic change exists in coding structure by shifting the majority of complexity to the decoder. The major task of exploiting the source redundancy is placed in the decoder. From those principles, the distributed vide o coding was derived. Recently, the WZ coding schemes [3, 4] including a SW (Slepian-Wolf) codec have been presented as the practical distributed video coding schemes. For the representative applications, such as wireless sensor networks, mobile camera phones, and wireless video surveillance systems, many small and lo w-power wireless sensors are employed. By observing the characteristics of the encoding source, it can be found that some region, such as background and area with smooth motion, can easily be reconstructed with side information without significant sacrifice of the quality of the reconstructed frame. Nowadays, few algorithms based on the coding mode are proposed to exploit both temporal and spatial redundancy to achieve higher coding efficiency. In [5], the coding mode is used to facilitate the motion compensation for the side information. In [6], a block classification is adopted in the codec with a simple skip mode in the encoder. As in [7], a similar mode decision scheme is included in the coding process either at the encoder or at the decoder so that intra coding is preferred when estimated correlation is weak for the coding block. These algorithms efficiently improve the rate-distortion performance of the codec at the price of the slightly increasing computational complexity. In this paper, two innovative components are proposed. First, the skip mode is adopted for the blocks with low motion. Second, the non-skip blocks are classified into two sets according to the motion activity so as to be encoded discriminately. The information for the skip blocks and non-skip blocks partition is derived at the decoder with various block sizes. For clarity of presentation, the rest of the paper is organized as follows. In section 2, a theoretical analysis for the proposed scheme is shown. In Section 3, the proposed scheme is described in details. A summary of the coding procedure for the proposed scheme is also present in this section. In Section 4, several experiments are performed in order to evaluate and compare the coding efficiency of the proposed approach, and finally, in Section 5, the conclusions are presented. This work was supported NSF of China under grant 60532060, 60702058, and CAST 200746.