Utilizing Median Filter as Pre-processing for Depth Map Compression

Abstract

Video codings are widely explored, and many standards have been established from multimedia industries. The 3D video format consists of the conventional 2D video format and depth maps.In order to support 3D video, depth video compression is needed to reveal geometric information for synthesis virtual views. Conventional video coding is not recommended directly employed into original depth video because its characteristics which can influence in compression ratio and synthesis performance. Depth information is able to be used for some image processing applications such as identification, segmentation and so on. Generally, the damage occurs in original depth video begins from acquisition process of depth video. This depth video is needed to be compressed for efficient network transmission. In this research, will be shown utilizing median filter as pre-processing method. Then, the output of pre-processing is fed into H.264/AVC encoding tool. From experimental results, utilizing median filter as a pre-processing technique already enrich the quality of video compression compared wihout using pre-processing at all.