Three-dimensional human face recognition method and system, and data processing device applying same

Abstract

The invention discloses a three-dimensional human face recognition method and system, and a data processing device applying the same. According to the embodiment of the invention, a series of key points having relatively high generic relevance are detected on a human face, and obvious difference exists when the key points are distributed on different human faces; and on the basis, fusion recognition of the three-dimensional human face is achieved through local characteristic matching and point cloud registration. Firstly, the unfavorable influence of expression change on human face recognition can be effectively reduced by using local characteristics; and secondly, the fusion recognition is achieved through characteristic matching and point cloud registration by comprehensively using local and global information of the human face, and the accuracy of the recognition method is further improved. Compared with the traditional method, the recognition method has the advantages that the recognition accuracy and the robustness are improved greatly.

Description

technical field [0001] The invention relates to the fields of computer vision and image processing, in particular to a method and system for three-dimensional face recognition based on local features and a data processing device using the same. Background technique [0002] Face recognition has received a lot of research in recent decades and achieved rich results. Most of the existing research focuses on face recognition based on two-dimensional images, but two-dimensional images are poorly adaptable to changes in illumination, posture, expression, age, etc. With the gradual promotion of point cloud acquisition equipment (that is, 3D image acquisition equipment), more and more researches have begun to turn to 3D face recognition based on point cloud. 3D face data has richer shape information in terms of information volume, and the acquisition of 3D data is not affected by light. However, 3D face recognition also has certain problems. 3D data matching at high resolution re...

AI Technical Summary

Problems solved by technology

Among them, the face recognition method based on global features mainly uses the overall attributes of the face to describe. It has high recognition accuracy for normal facial expressions, but for faces with changing expressions, the recognition accuracy is not ideal; face recognition based on local features The method mainly recognizes the details of the face, which has certain robustness to the change of expression, but due to the lack of global structural information, it is difficult to meet the accuracy requirements of face recognition.

Although some new methods unify global features and local features, the segmentation of face sub-regions relies on manual posture correction, which is less practical.

Images