基于支持向量机的手势识别研究  

Research on Gesture Recognition Based on Support Vector Machine

作　　者：洪期望 李捍东[1] HONG Qiwang;LI Handong(The Electrical Engineering College,Guizhou University,Guiyang 550025,China)

机构地区：[1]贵州大学电气工程学院,贵阳550025

出　　处：《微处理机》2022年第2期47-50,共4页Microprocessors

基　　金：国家自然科学基金(61663005)。

摘　　要：针对传统神经网络算法普遍存在识别准别率不高、运算量较大的问题,以手势分类识别为目标,通过人体肤色特征和SVM模型,设计一种手势识别检测模型。方法采用椭圆傅里叶算子算法提取出手势区域的轮廓,构成手势的特征向量。将肤色空间从RGB空间转到HSV空间下,从背景中将手势区域分离出来,在手势完整性方面引入形态学处理技术,有效填补手势图片中的黑洞区域和去除白点区域,直接对手势图片进行边缘处理。利用Qt制作客户端实现了基本数字手势的快速识别,并进行验证实验。实验结果表明,该方法在手势识别的准确率方面相比于传统算法都有所提高。Aiming at the problems of low recognition accuracy and large computation in traditional neural network algorithms, a gesture recognition and detection model is designed based on human skin color features and SVM model. The Elliptic Fourier operator algorithm is used to extract the outline of gesture area, which constituted the feature vector of gesture. The skin color space is transferred from RGB space to HSV space, the gesture area is separated from the background, and morphological processing technology is introduced in the aspect of gesture integrity to effectively fill the black hole area and remove the white point area in the gesture picture, and the gesture picture is directly edge processed. Using Qt to make the client realize the rapid recognition of basic digital gestures, and carry out verification experiments.Experimental results show that the method has improved the accuracy of gesture recognition compared with traditional algorithms.

关 键 词：手势识别 OpenCV库 支持向量机 Qt开发 傅里叶描述子 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]