Face recognition big data training method

Abstract

The invention discloses a face recognition big data training method, and belongs to the technical field of security and protection. The method comprises the following steps: placing a video acquisition device and a terahertz device at the same visual angle; sending the collected terahertz image and figure video stream to a server for preprocessing; importing the data processed in the step 200 into a big data processing unit for training; and storing the trained model into the processing unit of a neural network model. According to the invention, terahertz, a high-definition camera and a depth camera are adopted to construct multi-modal human body data and fuse and recombine the multi-modal human body data, the fused image can avoid the problems of low resolution and overlarge white noise of a terahertz image, and meanwhile, key points of bones can reflect rough information of a human body to meet the identification of dangerous object parts. According to the method, image information is greatly reduced, character main body information in the terahertz image is replaced with a skeleton reconstruction mode, the accuracy of an algorithm is greatly improved, and unattended recognition can be achieved when the method is applied to security check.

Description

technical field [0001] The invention belongs to the field of security technology, in particular to a large data training method for face recognition. Background technique [0002] The mainstream method of human body security inspection is mandatory inspection by security personnel holding security inspection devices or setting up security inspection doors and security inspection channels in important places such as airports. With the development of image technology, integrated security inspection systems such as face recognition security gates and security inspection face recognition systems have gradually been put into use, which has greatly improved the speed and accuracy of security inspections and realized the functional upgrade of the security inspection system. However, the electromagnetic field, X-ray and X-ray backscattering technologies used in traditional hand-held security detectors and security doors are far from meeting the above requirements because their worki...

AI Technical Summary

Problems solved by technology

According to the analysis of the applicant based on the above situation, the reasons for the above problems lie in the following aspects: 1) At present, the research and development of security inspection equipment is carried out in the way of face + terahertz recognition, instead of using data fusion of face images and terahertz images; 2) The resolution of the terahertz image is low, and the white noise is too large, which leads to the difficulty of direct fusion of multi-mode images when the complete terahertz image is used for big data analysis

Images