摘要
为了能够在图像采集区域将人的脸部和手势准确区分,需要将人机交互相互联系,同时也成为该领域的研究热点。脸部的特征和手的纹理高度相似,对手势准确识别产生了较强的干扰性。针对人脸与手会出现重叠的现象,本文基于人机交互手势识别特征,提出了边缘修复算法,在像素点空间匹配算法的基础之上,通过边缘修复算法修复人脸与手重合交界处的模糊边缘处,研究结果显示,边缘修复算法的正确率高达94.6%,而其他算法的正确率仅为8.7%,该算法对头部运动和手势变换频繁等情况具有显著的分辨效果。
In order to accurately distinguish peoples faces and gestures in the image acquisition area,it is necessary to connect human-computer interaction with each other,and it has also become a research hot spot in this field.The features of the face and the texture of the hand are highly similar,which produces a strong interference to the accurate recognition of the gesture.In view of the overlapping phenomenon between face and hand,this paper proposes an edge repair algorithm based on human-computer interaction gesture recognition feature.Based on the pixel point space matching algorithm,the edge repair algorithm is used to repair the intersection of face and hand.At the fuzzy edge,the research results show that the correct rate of the edge repair algorithm is as high as 94.6%,while the correct rate of other algorithms is only 8.7%.The algorithm has significant resolution effect on the head movement and gesture change frequently.
作者
梁英
Liang Ying(Xi'an Railway Vocational and Technical Institute,Xi'an,Shaanxi 710026,China)
出处
《西安轨道交通职业教育研究》
2020年第3期29-32,共4页
Xi'an Rail Transit Vocational Education Research
关键词
人机交互
手势特征
图像采集
边缘修复
Human Interaction
Gesture Feature
Graphic Gathering
Edge Repair
