| Summary: | 碩士 === 國立中央大學 === 電機工程研究所 === 93 === Recently, the masses have brought up all kinds of multimedia services to more and more demands of digit video, such as video on demand, distance learning and video surveillance. In these applications, compressed digital video is a major component of the multimedia data. Those various multimedia services demand different bit-rates to adapt network limitations and diverse terminal constraints decided by user. Among these techniques, video transcoding has allowed user to convert a previously compressed bit-stream into another format to meet various multimedia services.
In this thesis, we address in the topic of spatial-downscaling video transcoder in DCT domain. The proposed techniques include the Hierarchical Fast Motion Resampling (HFMR) with accuracy motion resampling, the Fast Refinement for Non-Integral-MV (FRNI) and the Dynamic Regulating Search (DRS) with low complexity motion vector refinement. Two kinds of motion vector refinement algorithms are design for different architectures and applications. Based on brute-force MC-DCT (Motion Compensation in DCT-Domain), FRNI can provide better quality than non-refine motion vector and reduce the complexity. DRS can utilize the filter for half-pixel motion vector in MC-DCT and efficient method for extracting MC-DCT block to improve the performance further.
From the experiments we have conducted, we believe that our proposed algorithms can improve the entire quality and also reduce the complexity for DCT-Domain video transcoder.
|
|---|
