A screening method and system for face pictures

Abstract

The present invention provides a screening method and system for human face pictures, including extracting the first picture containing the human face, enlarging the area containing the position of the human face in the first picture to generate a second picture, and performing grayscale processing to generate The third picture; using the edge detection algorithm to calculate the sharpness of the third picture, determining the first score based on the sharpness, and filtering pictures whose first score is less than the first threshold; determining the second score based on the face angle value, and filtering the first score A picture whose binary score is less than the second threshold; determine the third score based on the number of eyes of the face; determine the fourth score based on the pixel value of the first picture, the proportion of the face in the second picture, and the proportion of black, white and gray pixels, The comprehensive score value of the picture is obtained based on the first score, the second score, the third score and the fourth score, and the pictures whose comprehensive score is greater than the comprehensive threshold are output. The invention can filter a large amount of invalid data, reduce system pressure, and improve face recognition efficiency.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a screening method and system for face pictures. Background technique [0002] With the continuous development of deep learning and artificial intelligence technology, face recognition systems are more and more used in security, smart new retail, finance, subways, hotels, airports and other scenarios that require identity verification, such as remote bank account opening, wireless Renchao merchants automatically complete payment by swiping face, access control system, subway face swiping payment, airport automatic verification of witnesses, etc. [0003] Most of the existing face recognition is to directly obtain the captured faces from the camera for face recognition. The directly captured pictures contain many faces with motion blur or unclear focus, and many faces have inappropriate angles, scales and lighting. , which greatly affects the recognition accuracy, and also caus...

AI Technical Summary

Problems solved by technology

The disadvantage is that when evaluating the quality of portraits through some artificially specified dimensions, these dimensions may not have anything to do with the accuracy of face recognition, such as the proportion of portraits in the picture; on the other hand, it will be incomplete and often miss some Dimensions, such as occlusion between portraits, etc.

A large number of face pictures are required for training, and the training pictures need to be marked. The training work is complicated and the cycle is long, and it is easily affected by actual combat habits.

Images