平面赤足迹特征分析与身份识别方法

Feature Analysis & Identity Recognition of Planar Barefoot Impression

在对平面墨拓赤足迹进行原始特征分析的基础上,建立基于贝叶斯决策的足迹身份识别模型.在统一的坐标系下,对足迹进行特征点标注和轮廓分段,运用形状分析理论提取了基于角度和距离、基于区域和基于轮廓的3类共41个特征参数。经过特征选择,利用贝叶斯决策进行身份识别,以266人的532枚足迹作为实验数据,正确识别率达到97.8%。实验结果表明,在同等条件下,平面墨拓赤足迹的形态特征具有身份可识别性。

Based on a through analysis of planar barefoot impressions, a footprint recognition model based on Bayesian decision was established. Landmarks location and contours segmentation of footprint were performed under a unified coordinate system. 41 of three different kinds of features, i. e. , features about angles and distances, about regions, and about contours, were extracted. After feature selection, footprint verification was performed using the Bayesian decision theory. 532 footprints of 266 persons were used in our experiment, and the correct verification rate was as high as 97.8%. The result shows that planar barefoot impression can be used for identification.