FAST CUMULANT-MATCHING ALGORITHMS FOR IMAGE MOTION ESTIMATION
碩士 === 大同工學院 === 電機工程學系 === 84 === A class of algorithms is presented that estimates the motion vector from two successive image frames consisting of non- Gaussian signal plus Gaussiannoise. The cumulant-matching algorithm is developed th...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
1996
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/11224741256348782164 |
| Summary: | 碩士 === 大同工學院 === 電機工程學系 === 84 === A class of algorithms is presented that estimates the motion
vector from two successive image frames consisting of non-
Gaussian signal plus Gaussiannoise. The cumulant-matching
algorithm is developed that follows from a cumulant-error
optimality criterion. This exhaustive search algorithms
needhuge amounts of computations and data manipulations. In
this thesis, wepropose two techniques to reduce the
computational complexity by hierarchicalsearch and new cumulant
based criterion. Experimental results show that combining these
techniques provides high-speed and high-precision
mationestimators.
|
|---|