Low complexity H.263 to H.264 video transcoding using motion vector decomposition

The H.264 adopts various block types and multiple reference frames for motion compensation. For transcoding a video sequence from the H.263 format to the H.264 format, it is beneficial to reuse as much information as possible in the original sequence. It is always a good idea to design a video transcoder which aims at achieving a transcoded video sequence with good quality and lower bitrate, and without introducing high computational complexity during the transcoding process. Motivated by this, an effective motion vector decomposition algorithm is proposed for transcoding videos from the H.263 format to the H.264 format. The algorithm concentrates on macroblocks which consume more bits in the H.263 video sequence. In other words, a non-optimal motion vector is splitted into smaller sizes since the H.264 support various block types. Using the proposed prediction error measure and diversity of the motion vector measure, the motion vector can be decomposed adaptively and the amount of th...

• Type: Conference
• Year: 2005
• Where: ISCAS
• Authors: Kai-Tat Fung, Wan-Chi Siu
• Keywords: ISCAS 2005, Motion Vector, Video Sequence, Video Transcoder