Research On The Discrimination Method Of Fuzzy Face Images Based On High Frequency Information

The deterioration of image quality caused by blurring is one of the most important factors effecting face recognition accuracy.The discrimination of blurring face can provide useful guidance for face image's acquisition,processing and recognition,and thereby reduce the adverse impact caused by blurring on face recognition.Most of the existing researches on the blurring face discrimination are mostly based on the general research situation of the blurring image.The method is simple and the accuracy is not high,so it is lack of research on the discrimination of blurring face images.It does not make good use of the easy-to-obtain and stable high-frequency information features to discriminate the blurring images.At the same time,it does not effectively combine the face features that have a direct relationship with the blurring of face images.According to the present situation of research on blurring face image,through careful analysis and reasonable use of the high frequency information,this paper put forward three methods of judging blurring face image:1)a method of blurring face discrimination based on analysing the high-frequency information threshold of the whole image,2)a method of blurring face discrimination based on analysing the high-frequency information threshold of feature region,3)a method of blurring face discrimination based on training the high-frequency information classifier of feature region.A large number of experiments were carried out on the public data set,and the validity of the proposed blurring face discrimination based on high-frequency information analysis was proved by comparing the experimental results with the previous research methods and the three methods proposed in this paper.The three methods proposed in this paper,from simplicity to complexity,are a gradually improving process.At the same time,the corresponding discriminant accuracy is also gradually improved.Through the comprehensive analysis of the above three methods and the other two blurring face discrimination methods briefly introduced in this paper,we can get the conclusion that the high-frequency information of the image is easy to obtain and has a more direct relationship with the image blur,and to analyze the high-frequency information and extract appropriate features can effectively discriminate the blurred face image.Combined with the extraction of face landmarks to achieve a detailed analysis of image's blur can further improve the discrimination accuracy.In the process of discrimination,the classifier can be introduced reasonably,and combined with the excellent feature classification ability of the classifier,the discrimination accuracy can be improved again.