In this paper a novel Hierarchical Sum of Double Difference metric, HSDD, was introduced. It was shown, as opposed to conventional Sum of Absolute Difference (SAD) metric, how this embedded-coding aware metric can jointly constrain the motion vector searching in both temporal and spatial (quad-tree) directions under multiresolution motion estimation (MRME) framework. The temporal-spatial co-optimization concept from HSDD brings us the motion compensation pyramid with better shape. The reward is that fewer bits are spent later for describing the isolated zeros. The compression performance of HSDD easily exceeds the performance of its competitors, especially when high compression ratios are used.
Identifer | oai:union.ndltd.org:NSYSU/oai:NSYSU:etd-0709102-144112 |
Date | 09 July 2002 |
Creators | Hsu, Chin-Hsun |
Contributors | Yan-Hwang Kwu, Chu-Sing Yang, Lih-Shyang Chen, Wen-Shyong Hsieh |
Publisher | NSYSU |
Source Sets | NSYSU Electronic Thesis and Dissertation Archive |
Language | Cholon |
Detected Language | English |
Type | text |
Format | application/pdf |
Source | http://etd.lib.nsysu.edu.tw/ETD-db/ETD-search/view_etd?URN=etd-0709102-144112 |
Rights | unrestricted, Copyright information available at source archive |
Page generated in 0.0021 seconds