Three-dimensional face reconstruction method, system and device for fine structure

Abstract

The invention belongs to the technical field of image processing and pattern recognition, particularly relates to a three-dimensional face reconstruction method, system and device for a fine structure, and aims to solve the problem of poor three-dimensional face reconstruction precision. The method comprises the steps of obtaining a to-be-reconstructed two-dimensional face image; obtaining a three-dimensional space transformation function and an initial three-dimensional face shape; performing spatial transformation on the initial three-dimensional face shape, and mapping each pixel of an image face region to a UV texture space of a 3DMM model to obtain a UV texture map; obtaining a UV visible image and extracting features to obtain an attention feature map; mapping each point of the initial three-dimensional face shape to a UV texture space to obtain a UV shape graph; multiplying the attention feature map and the UV texture map, and adding the multiplied attention feature map and UV texture map to the UV shape map; and obtaining the update amount of each point of the 3DMM face model, and adding the update amount to each point corresponding to the initial three-dimensional face shape to obtain a three-dimensional reconstruction result. According to the invention, the precision of face model reconstruction is improved.

Description

technical field [0001] The invention belongs to the technical field of image processing and pattern recognition, and in particular relates to a fine structure-oriented three-dimensional face reconstruction method, system and device. Background technique [0002] Most of the current 3D face reconstruction algorithms are based on 3D variable model (3DMM) for shape reconstruction. However, the construction of most of these face models only uses a few hundred scanned point clouds, the age span and ethnicity span are very small, and these images are often taken in a controlled environment, with the face on the front and The expression is natural. Such a small training set is far from enough to describe all the changes that may exist in the actual face, so these models often perform poorly on pictures with large changes in expression or posture. [0003] The existing mainstream method is to reconstruct two-dimensional face images based on convolutional neural network, which gene...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the above problems in the prior art, that is, in order to solve the problems of lack of details and strong sense of model in the existing 3D face reconstruction based on the 3DMM model, resulting in poor reconstruction accuracy, the present invention proposes a fine structure-oriented 3D A face reconstruction method, the method comprising:

Images