Dual-camera face identification method and system

Abstract

The present invention discloses a dual-camera face identification method and system. The method comprises: collecting an original black-and-white image and an original color image of a face through a first camera and a second camera, and sending the original black-and-white image and the original color image to a processor; reading the position deviation information stored in the EPROM through the processor, performing displacement image processing of the collected original black-and-white image according to the reading position deviation information, and adding a relative position figured frame in the black-and-white image after displacement; performing balance preprocessing of the collected original color image through the processor; performing feature extraction of the black-and-white image after displacement and the color image after the balance preprocessing, and obtaining the face feature information; and calculating the similarity of the obtained face feature information and face image feature information stored in advance through the processor, and obtaining a final face identification result. The dual-camera face identification method and system have small error and high identification rate, and can be widely applied in the face identification field.

Description

technical field [0001] The invention relates to the field of face recognition, in particular to a dual-camera face recognition method and system. Background technique [0002] Existing face recognition equipment generally adopts dual cameras: one is a visible light camera and the other is an infrared camera. Infrared cameras are hardly affected by lighting, so infrared images are used for actual recognition. What is displayed on the screen of the face recognition device is not an infrared image, but a visible light image, because the infrared image is a black and white image that cannot be displayed on the screen. However, the two cameras are not installed in the same position (usually arranged left and right), so the two images captured by the actual face recognition device must have deviations. If the face recognition device does not consider this deviation during face recognition, although the user feels that he has looked at the camera correctly, the situation that the...

AI Technical Summary

Problems solved by technology

However, what is displayed on the screen of the face recognition device is not an infrared image, but a visible light image, because the infrared image is a black and white image that cannot be displayed on the screen.

However, the two cameras cannot be installed in the same position (usually arranged left and right), so the two images captured by the actual face recognition device must have deviations

If the face recognition device does not consider this deviation during face recognition, although the user feels that he has looked at the camera correctly, the situation that the face recognition device cannot recognize will appear, which will easily lead to excessive errors in the recognition results

[0003] In addition, most of the current face recognition devices only perform face recognition based on black and white images, while ignoring the recognition function of color images such as visible light images.

Images