Face angle and face ambiguity analysis method and system and computer equipment

Abstract

The invention discloses a face angle and face ambiguity analysis method and system, and computer equipment. The method comprises the steps of S1, neural network model training, S2, to-be-detected image extraction, S3, image face detection and correction, S4, face angle and ambiguity prediction, and S5, attribute output. Through a first neural network model based on deep learning, a face frame andface feature points in a picture can be rapidly detected, and a face frame position and a face feature point position are output; after correcting and amplifying the face frame, a corrected and standardized face image is intercepted; face angle and face ambiguity analysis are carried out based on the head picture on the captured corrected and standardized face picture through a second neural network model. The confidence coefficient of output face angle classification and the regression value of the face ambiguity are obtained and predicted, the problems of defocusing and motion ambiguity aresolved, high-quality pictures can be extracted, various data analysis, comparison and other operations can be carried out on the basis of the high-quality pictures, and the accuracy of subsequent analysis and prediction is improved.

Description

technical field [0001] The invention relates to the technical field of infrared living body detection of smart security terminals, in particular to a face angle and face ambiguity analysis method, system and computer equipment. Background technique [0002] At present, we can see all kinds of cameras everywhere in our daily life, most of which have storage functions, and generate a large amount of data every day. These data can be used for various video structural analysis, but there are various The camera is installed in different environments, and many problems such as overexposure, underexposure, out-of-focus caused by poor focus, and motion blur will cause uneven image quality, which will also cause trouble for later data detection and analysis. For example, in the all-in-one machine for automatic verification of personal identification, comparing a face image with poor imaging quality with an ID card image often leads to misjudgment of the result as someone who is not t...

AI Technical Summary

Problems solved by technology

[0002] At present, we can see all kinds of cameras everywhere in our daily life, most of which have storage functions, and generate a large amount of data every day. These data can be used for various video structural analysis, but there are various The camera is installed in different environments, and many problems such as overexposure, underexposure, out-of-focus caused by poor focus, and motion blur will cause uneven image quality, which will also cause trouble for later data detection and analysis. For example, in the all-in-one machine for automatic verification of personal identification, comparing a face image with poor imaging quality with an ID card image often leads to misjudgment of the result as someone who is not the person, not only does not increase the efficiency of access control security detection, but on the contrary Caused blockage, therefore, in order to solve the above-mentioned problem, the inventor proposes a kind of face angle and face ambiguity analysis method based on head photo

Images