基于CNN的手势识别技术研究  

Research on gesture recognition based on CNN

作　　者：叶茂华[1] 

机构地区：[1]淮安信息职业技术学院,江苏淮安223003

出　　处：《无线互联科技》2017年第20期129-130,共2页Wireless Internet Technology

摘　　要：手势操作作为一种全新的操作方式,在智能设备中得到了广泛应用。传统的手势识别方法需要复杂的预处理过程,识别速度与准确度比较差。文章提出一种基于CNN的手势识别技术,取得了较好的识别效果,识别速度也有明显的提升。Gesture manipulation, as a new way of operation, has been widely applied in intelligent devices. Traditional gesture recognition methods generally need complex preprocessing with poor recognition speed and accuracy compared with the new. In this paper, a gesture recognition method based on convolutional neural network is proposed, good recognition results are achieved with recognition speed improved obviously.

关 键 词：卷积神经网络 手势识别 特征提取 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]