Feature-based three-dimensional face modeling method

Abstract

The invention discloses a feature-based three-dimensional face modeling method. The method comprises: (1), a face is divided into a plurality of regions; (2) a feature point of each region is extracted and a facial form in each region is adjusted locally by comparing a preset feature point of a standard model; (3), adjusted two-dimensional feature points are mapped to a three-dimensional face model from two-dimensional coordinates; (4), after completion of coordinate mapping, the feature points are processed by using a B-spline function to realize fitting of a B-spline curve; and (5), all facial regions are spliced by using a smooth transition algorithm. A facial model generated based on the feature-based three-dimensional face modeling method has a high reduction degree and is vivid in details of the skeleton, contour, and the five sense organs; and thus problems that the previous face modeling method is lack of a three-dimensional effect and the facial stiffness is high can be solved.

Description

technical field [0001] The invention relates to the technical field of three-dimensional human body modeling, in particular to a feature-based three-dimensional face modeling method. Background technique [0002] The traditional 3D face modeling is to use a set of mesh face models composed of triangular faces, to compare and adjust each triangular face, and then use the smooth transition algorithm to polish the face model. The problem with this solution is that The granularity of the adjustment is too high, because the number of triangles in the entire model is fixed, so it is bound to appear that some areas cannot be adjusted to the desired pattern. Moreover, during the adjustment process of the model, what it adjusts is a piece of individual triangular faces. In some cases, the structure of the entire face may be cut into pieces. Even if the smooth transition algorithm is used, it is difficult to polish a complete and recognizable person. face model. Contents of the inv...

AI Technical Summary

Problems solved by technology

[0002] The traditional 3D face modeling is to use a set of mesh face models composed of triangular faces, to compare and adjust each triangular face, and then use the smooth transition algorithm to polish the face model. The problem with this solution is that The granularity of the adjustment is too high, because the number of triangular faces in the entire model is fixed, so it is bound to appear that some areas cannot be adjusted to the desired pattern.

Moreover, during the adjustment process of the model, what it adjusts is a piece of individual triangular faces. In some cases, the structure of the entire face may be cut into pieces. Even if the smooth transition algorithm is used, it is difficult to polish a complete and recognizable person. face model

Images