Three-dimensional face identification method for safe case

Abstract

The invention discloses a three-dimensional face identification method for a safe case. The method comprises the following steps: (1) adjusting the relative positions between left and right cameras and the face to finish the acquisition of two-dimensional images of the face; (2) obtaining a three-dimensional coordinate of key feature points of the face by combining the inner and outer parameters of the cameras; (4) using the three-dimensional coordinate in step (3) to calculate distance matrixes of the feature points, so as to extract the geometric features of the three-dimensional face; and (5) identifying by a counter-propagation neural network. In the identifying process, according to the method provided by the invention, the automatic positioning of two-dimensional feature points can be carried out on the two-dimensional image by an initiative shape model, and the three-dimensional coordinate of the feature points can be obtained by combining the inner and outer parameters of the cameras, therefore, the complicate three-dimensional rebuilding of the face can be prevented; furthermore, the correct recognition rate can be improved by a way of identifying through the counter-propagation neural network.

Description

technical field [0001] The invention belongs to the technical field of face recognition, and in particular relates to a three-dimensional face recognition method used in safes. Background technique [0002] Face recognition is easily affected by factors such as environment, posture, and expression. Compared with two-dimensional face image data, 3D face image data can provide more complete and richer recognition information. And the requirements of factors such as facial expressions are strong. [0003] The current face recognition methods mainly tend to recognize methods based on 3D face models. Although theoretically, 3D face recognition methods have better performance than 2D face recognition methods, in practical applications, 3D face recognition methods are still in the advanced stage. In the immature state, there are mainly the following difficulties: 1) It is difficult to obtain 3D data. Although laser and structured light acquisition equipment can obtain more accurat...

AI Technical Summary

Problems solved by technology

[0003] The current face recognition methods mainly tend to recognize methods based on 3D face models. Although theoretically, 3D face recognition methods have better performance than 2D face recognition methods, in practical applications, 3D face recognition methods are still in the advanced stage. In the immature state, there are mainly the following difficulties: 1) It is difficult to obtain 3D data. Although laser and structured light acquisition equipment can obtain more accurate data, they are expensive and unacceptable to general users, and structured light acquisition is easily affected by the environment; 2) due to The traditional 3D face recognition method is based on the 3D reconstruction of the face, and the computational complexity of 3D reconstruction is large, which makes the traditional 3D face recognition method complex.