An Optimal Frame Selection Method Based on Dynamic Face

Abstract

The invention discloses an optimal frame selection method based on a dynamic face. The method comprises the following steps: S101, video sample collection: collecting video sample information in a front-end camera in advance; S103, extracting target group information: extracting continuous frame pictures of each target person from the video sample information obtained in the step S101 as target group units, and storing the target group units; S105, extracting target group feature information: extracting face information from the target group units obtained in the step S103 by using MTCNN, andstoring the extracted face information as the target group feature information; and S107, information grouping: carrying out manual labeling on the target group feature information obtained in the step S105, and carrying out manual labeling according to the image quality of the face information. According to the invention, through the mutual cooperation, respective doing and mutual cooperation ofall the above functions, the selection of the dynamic face optimal frame is rapidly and effectively completed, the manpower and material resource cost is saved in the practice of dynamic face recognition, and the accuracy is significantly improved.

Description

technical field [0001] The invention relates to the technical field of image detection and intelligent recognition, in particular to a method for selecting an optimal frame based on a dynamic human face. Background technique [0002] Face detection is a very important and practical research topic in today's visual field. It is used in various fields in real life, such as public security, finance, network security, property management, and attendance. [0003] Face recognition currently mainly includes static face recognition and dynamic face recognition. Static face recognition is to recognize within a specific area or range, that is to say, the recognition has relatively high requirements for angle, distance, and position. The feature of static face recognition is that the user capacity is small, which is more suitable for the attendance of some small companies. Because it is static, the price is relatively cheap. The graphic recognition rate is high, which can reach mor...

AI Technical Summary

Problems solved by technology

[0005] 1. Illumination problems, such as side light, top light, backlight and high light may appear, and the light may be different in different periods, and even the light in each position in the monitoring area is different

[0006] 2. Faces with various poses and ornaments

[0007] 3. The image quality of the camera is uneven

[0008] 4. Frame loss and face loss

Images