Three-dimensional face reconstruction method based on gray scale and depth information

Abstract

The invention discloses a three-dimensional face reconstruction method based on gray scale and depth information. The three-dimensional face reconstruction method includes the steps of identifying face gray scale information, identifying face depth information, conducting multimodal face identification, matching by a 3D model and conducting face 3D reconstruction. Feature areas of face data are located, and registering and feature extraction are carried out through feature points. Adaboost algorithm is used to select the most effective feature for classification. A nearest neighbor classifier is used to calculate matching scores to achieve multimodal face identification. Finally, face reconstruction is completed by matching the local 3D model, the. The performance and efficiency of a face recognition system are improved through a fusion strategy. 3D cascade regression is used, dense 3D point set is selected, and a face is completely marked to avoid the logo position change. The problems of inconsistency of action change position points and self-closure are solved, and the calculation cost is reduced. The method is highly universal, and the real-time effect is good.

Description

technical field [0001] The invention relates to the technical field of face recognition, in particular to a three-dimensional face reconstruction method based on gray scale and depth information. Background technique [0002] The 3D face mesh reconstruction method can be used for criminal monitoring, face reconstruction without criminal fingerprints or identity information, can also be used for 3D printing, and can also be used for 3D face modeling and animation In fields such as production, it has a great influence on various fields. Compared with 2D face recognition, 3D face recognition has the advantages of being robust to illumination and less affected by factors such as posture and expression. Therefore, after the rapid development of 3D data acquisition technology and the great improvement in the quality and accuracy of 3D data, Many scholars devote their research to this field. [0003] Face grayscale images are easily affected by illumination changes, while face de...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problem that the face grayscale image is easily affected by illumination changes, and the face depth image is easily affected by data acquisition accuracy and expression changes, the purpose of the present invention is to provide a three-dimensional face reconstruction based on grayscale and depth information Method, by fusing grayscale and depth information to obtain a multimodal system, by collecting two-dimensional grayscale information and three-dimensional depth information, using the collected information points to reconstruct the face shape by matching the local 3D model

Images