Robust Recognition Using Eigenimages

详细信息    查看全文

• 作者：Leonardis ; Ale&scaron ; Bischof ; Horst
• 关键词：appearance-based matching ; principal component analysis ; robust estimation ; occlusion ; minimum description length
• 刊名：Computer Vision and Image Understanding
• 出版年：2000
• 出版时间：April, 2000
• 年：2000
• 卷：78
• 期：1
• 页码：99-118
• 全文大小：256 K

文摘

The basic limitations of the standard appearance-based matching methods using eigenimages are nonrobust estimation of coefficients and inability to cope with problems related to outliers, occlusions, and varying background. In this paper we present a new approach which successfully solves these problems. The major novelty of our approach lies in the way the coefficients of the eigenimages are determined. Instead of computing the coefficients by a projection of the data onto the eigenimages, we extract them by a robust hypothesize-and-test paradigm using subsets of image points. Competing hypotheses are then subject to a selection procedure based on the Minimum Description Length principle. The approach enables us not only to reject outliers and to deal with occlusions but also to simultaneously use multiple classes of eigenimages.