Study of Multiple Hexagon Search Algorithm for Real-Time Object Tracking System

• Main Authors: Chia-hsiu Ho, 何嘉修
• Other Authors: Chao-yin Hsiao
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/68516979358021442612

碩士 === 逢甲大學 === 材料與製造工程所 === 95 === Block-matching algorithm (BMA) has been applied on image tracking system very extensively. In BMA, the search speed and the computation accuracy on object motion estimation are decided by which search patterns are used in the algorithm. In order to get the real time image tracking capability and reduce unnecessary data on the time domain of the video sequence, in this article, we used statistics technology to analyze motion vector (MV) probability, and got the MV’s characteristics to bring up the new search patterns and the flow process, and to promote the search efficiency. Compare this proposed method with the other similar search algorithms. It has as high as above 25% search efficient and it is much more robust in image quality.