Method for synchronously realizing face three-dimensional point cloud feature point positioning and face segmentation

Abstract

The invention belongs to the field of computer vision and the field of point cloud processing, and relates to a method for synchronously realizing human face three-dimensional point cloud feature point positioning and human face segmentation, which comprises the following steps: S1, point cloud initialization: inputting human face point cloud data; S2, projection: projecting the point cloud information with the texture onto a 2D image; S3, 2D feature point positioning: positioning face feature points on the projected 2D image; S4S4, solving 3D feature points: solving the feature points of theface three-dimensional point cloud according to the corresponding relationship; S5, segmentation: cutting the face point cloud data by using the feature point information; S6, trimming: removing pointcloud outliers; S7, iteration: returning to S2 to resolve the point cloud feature points, and performing iteration until the point cloud feature points are stable; and S8, outputting: outputting theclipped face point cloud and the clipped face 3D feature points. According to the method, the 3D human face feature points are synchronously positioned, the human face point cloud is cut, the two processes are mutually promoted, the human face point cloud feature points can be solved with high precision, and the method is simple, convenient, feasible, efficient and practical.

Description

technical field

[0001] The invention belongs to the fields of computer vision and point cloud processing, and in particular relates to a method for synchronously realizing feature point positioning and face segmentation of a three-dimensional point cloud of a human face. Background technique

[0002] With the development of 3D scanning technology, 3D scanning of human face is widely used. Face segmentation and accurate 3D feature point positioning of face 3D point cloud data play an important role in face alignment, face recognition, face 3D printing, face animation, virtual reality and other fields. The cropping methods of face point cloud data mainly include the following methods:

[0003] (1) Random sampling consensus algorithm (RANSAC) algorithm. This algorithm can estimate a mathematical model from a set of scanning data containing "outlier points" by randomly cutting off some existing data and iterating multiple times. This algorithm is suitable for finding the expr...

Images