Abstract: This paper proposes a gait recognition algorithm based on Fuzzy Principal Component Analysis(FPCA) for gait energy image. It preprocesses the original gait sequence to obtain Gait Energy Image(GEI), extracts the eigenvalues and eigenvectors by FPCA which are called fuzzy components, projects the eigenvectors into lower dimension space, and utilizes the NN classifier in feature classification. The algorithm is tested on CASIA database, and experimental results show that compared with other algorithms, its recognition performance is better.

Key words: Gait Energy Image(GEI), Fuzzy Principal Component Analysis(FPCA), gait recognition, feature extraction, objective function

摘要： 提出针对步态能量图的基于模糊主成分分析的步态识别算法。通过对原始步态序列进行预处理得到步态能量图，利用模糊主成分分析提取出特征值和对应的特征向量，获得模糊主成分后将其映射到低维空间，并使用最近邻法进行分类。在CASIA数据库上对算法进行验证，实验结果证明，该算法与同类算法相比具有更好的识别性能。

关键词: 步态能量图, 模糊主成分分析, 步态识别, 特征提取, 目标函数

CLC Number: 

• TP391.41