Method and device for reconstructing three-dimensional face based on single face image

Abstract

The invention discloses a method and a device for reconstructing a three-dimensional face based on a single face image. The method comprises the following steps of: performing posture recognition on a face image by utilizing face structure priori knowledge, and estimating rotating direction and angle of a face plane by combining face measurement and projective geometry so as to set a rotating angle of the three-dimensional face; estimating the depth of two-dimensional characteristic points on the face image by adopting an artificial neural network, and acquiring three-dimensional coordinates of the characteristic points; converting a common three-dimensional face model into a specific model by adopting a Dirichlet free deformation algorithm; and mapping the three-dimensional face model by adopting characteristic points extracted from the two-dimensional image. Therefore, the true three-dimensional model is constructed through the single face image, the three-dimensional modeling accuracy is effectively improved, the three-dimensional modeling time is shortened and the three-dimensional modeling cost is reduced.

Description

technical field [0001] The present invention relates to the technical fields of digital image processing, machine vision and artificial intelligence, in particular to a method and device for realizing three-dimensional face reconstruction based on a single face image. Background technique [0002] The current image-based face modeling methods are mainly divided into two types: one is a reconstruction method based on the shape of multiple pictures or orthogonal pictures, and the other is a three-dimensional face modeling method based on a single image. [0003] The traditional shape reconstruction method based on multiple photos generally uses hardware (such as special equipment such as 3D scanner) to obtain the 3D information of the face. Although the theoretical basis is relatively mature, its realization cost and implementation method are relatively difficult; Orthogonal image reconstruction method, since the frontal image is orthogonal to the side image, the three-dimensi...

AI Technical Summary

Problems solved by technology

However, this method uses 200 scanned 3D face models, through linear combination, combined with 22 parameters such as rotation angle and illumination, and iterative optimization algorithm to obtain the target model, which takes too long

[0005] Moreover, these methods for 3D face modeling from a single image do not consider how to obtain the angle of the face and the depth of the feature points from an image, but only use the feature point coordinates of the 2D image, thus affecting the 3D facial features. modeling accuracy

Images