摘要
为了快速准确地进行人体运动步态识别,基于人体轮廓统一Hu矩的不变矩特征,提出了一种新的步态识别算法。该算法首先对每个序列进行运动轮廓提取,然后计算每个序列的不变矩,构成步态特征空间;在此基础上采用支持向量机训练步态,最后用支持向量机进行步态识别。实验结果表明,该算法具备快速、稳健的特征,取得了较好的识别率,初步具备了实际应用的价值。
In order to increase the accuracy and rate of gait recognition, an approach to improve the gait recognition is proposed based on unified Hu＇s invariant moments of outer contour. The vector data scanned from binarized silhouette of a walking person are chosen as the image feature. Then the multi-class support vector machine （SVM） models are trained by the decomposed feature vectors. The gaits are classified by the trained SVM models. Extensive experimental results demonstrate that the proposed algorithm performs at an encouraging recognition rate at relatively lower computational cost.
出处
《微计算机信息》
2010年第13期197-198,196,共3页
Control & Automation
关键词
步态识别
HU不变矩
支持向量机
gait recognition
Hu moment
support vector machine（SVM）
