摘要
采用极半径不变矩作为描绘子对平面单足足迹进行粗匹配实验。设计了"距离排序法"作为粗匹配的分类器。使用极半径不变矩的5维特征值作为描绘子识别实验结果为在排除83%前嫌人员的条件下,识别准确率为86%。实验的结果表明使用极半径不变矩作为描绘子进行的粗匹配比使用足长与足宽的粗匹配的识别准确率和前嫌人员排除率都更高。
The plane single-footprint was coarse matched by polar-radius invariant moment as descriptors in this paper. A method of distance sorting was designed as coarse match classifier. The recognition experimental result that the identifiable accuracy is 86 ~ 5D characteristics of polar-radius invariant moment when excluding rate of suspect is 83 %. Experimental results indicate the identifiable accuracy and the exeludeing rate by the method to coarse match single-foot print used polar-radius invariant moment as descriptors are higher than that used foot-length and foot-breadth as descriptors.
出处
《电子测量技术》
2009年第12期72-73,76,共3页
Electronic Measurement Technology
关键词
单足足迹识别
粗匹配
极半径不变矩
分类器
距离排序法
recognition for single-footprint
coarse match
polar radius invariant moment
classifier
method of distance sorting
