SHAPE MATCHING UTILIZING THE SIMILARITY MEASURE OF TARGET SHAPES

摘要：

This paper presents a new efficient method, based on the similarity measure of the target shapes (MSTS), for shape matching issue.MSTS uses improved GVF snake to obtain the contours and takes curve evolution to predigest the contour at first And then it operates the tangent space representation to deduct the predigested target shapes.Finally, the method on the base of the similarity measure of target shapes is adopted to match shapes.Experiments demonstrate that MSTS, combined with the improved GVF snake and curve evolution algorithms effectively, can express the target shapes simply and accurately, and also show a better balance among anti-jamming, the precision of matching and the speed of matching.

关键词： GVF snake Shape matching Curve evolution Tangent space representation

作者: GANG XU YAN-YAN FU

作者单位: School of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, China

会议类型: 国际会议

会议名称: 2007 International Conference on Machine Learning and Cybernetics(IEEE第六届机器学习与控制论国际会议)

会议地点: 香港

会议语种:英文

页码: 1515-1520

在线出版日期: 2007-08-19（万方平台首次上网日期，不代表论文的发表时间）