基于几何特征的静态字母手势识别算法

针对聋哑人哑语手势自动识别问题的复杂性,研究了手势几何特征的多样性及提取和识别方法,提出了一种基于几何特征的手势识别算法.首先,对手势图像进行肤色分割、边缘检测以及逻辑运算,然后,计算其质心面积等多项几何特征,通过实验方法测定最佳特征权值,最后,将其与样本图像特征值进行匹配,最佳匹配即为检测结果.根据30个字母手势创建了3套手势库,其中1套作为样本集,2套作为测试集.实验结果表明,通过该方法进行特征提取来识别汉语字母手势,可有效提高识别率,测试集识别率达到93.33％.

Algorithm of Gesture Recognition Based on Geometric Features

Considering of the complexity of sign language gesture automatic identification,an algorithm of gesture recognition based on geometric features is given with the research on the diversity of gesture＇s geometric features,extraction and identification methods.Firstly,process images of hand gesture by complexion detection,edge detection and logical operation.Then calculate multi-geometry features such as the center of mass and area,and get the feature weighting coefficients by experimental methods.Finally,match the image feature value to the sample,and the detection result is obtained as the best match.Three sets of gesture library are created by 30 alphabet gestures,and one is set as a sample set,while other two are set as test sets.Experiments show that the algorithm which uses feature extraction to recognize alphabet gesture can improve the recognition rate.