一种有遮挡人脸识别方法改进

An improved method for obstructing face recognition

随着人工智能、计算机视觉、模式识别近些年的快速发展,人脸识别已广泛应用于公共安全、验证系统、视频会议、人机交互等各个领域。正常的人脸识别,已经得到充足的发展。提高在非正常拍摄下的人脸识别就变得迫切需要,以特征提取方法和图像匹配技术为研究对象,将人脸识别作为应用背景,对不同的特征提取方法进行了讨论和研究,并改进了一种基于矩阵γ范数带遮挡的人脸识别的实验方法。通过选择合适的参数、控制不同结构型遮挡噪声大小来比较不同方法之间的差异。实验结果表明,本文的实验方法较其他方法在不同的遮挡情况下,可以获得更好的效果。

With the rapid development of artificial intelligence, computer vision, and pattern recognition in recent years, face recognition is used in various fields such as public safety, verification systems, video conferencing, and human-computer interaction. Normal face recognition has been developed to adequate development. Therefore, it is more necessary to improve face recognition under abnormal shooting. This paper takes feature extraction method and image matching technology as the research object, and uses face recognition as the application background to discuss different feature extraction methods. Researched and improved an experimental method for face recognition based on matrix γ norm masking. By selecting the appropriate parameters and controlling the size of the occlusion noise of different structures, the differences between different methods are compared. Experimental results show that the experimental method of this paper can obtain better results than other methods under different occlusion conditions.