低质量图像人脸图像模糊消除研究

Research on the Fuzzy Elimination of Face Image in Low Quality Image

目前,我国人脸识别技术在模型训练和计算机算法设计都是在质量好、辨识度高的图像的基础上进行人脸识别,但对于低质量、模糊度高的图像,识别效果较差。公安系统、支付系统、智能监控系统、安全监控系统等人脸图像来源不一,图像质量参差不齐,如光线效果差、分辨率低等因素导致图像模糊、质量低,增加了人脸图像识别的难度。基于此,笔者在现有人脸图像模糊消除技术不足的基础上研究了低质量图像人脸图像模糊消除方法。

At present,face recognition technology in our country is based on the image with good quality and high recognition degree in model training and computer algorithm design,but the effect of face recognition is poor for the image with low quality and high fuzzy degree,while the face image sources such as public security system,payment system,intelligent monitoring system and security monitoring system are different,and the effect of image quality is uneven Poor light effect,low resolution and other factors lead to image blur and low quality,which increases the difficulty of face image recognition.On the basis of the research on the shortcomings of the existing face image blur elimination technology,this paper puts forward the research on the low quality face image blur elimination.