Human face three-dimensional model acquiring method based on stereo matching

Abstract

The present invention relates to a human face three-dimensional model obtaining method based on solid matching. The method is realized by the steps as follows: firstly, two sets of vidicons are used for respectively shooting a picture of human face from two different angles at the same time, one picture is used as a reference image, and the other picture is used an aligning picture; secondly, two sets of vidicons are demarcated respectively; thirdly, the profile lines of a face, eyes, eyebrows, a nose and a mouth are picked up from the reference picture, all pixel points on the profile lines are matched by a dynamic planning arithmetic, and the human face in the reference picture is divided into seven areas of the eyes, the eyebrows, the nose and the mouth to be matched according to the matched profile lines; fourthly, the pixel points to be matched in the reference picture are matched from the external edge line of each area to be matched towards the inside of the area by the arithmetic combining area increasing with dynamic planning; fifthly, the three-dimensional coordinate of human face actual points corresponding with the matched points is calculated in terms of computer parameter, and the three-dimensional point cloud data of a whole human face is obtained further.

Description

Technical field: [0001] The invention belongs to the technical field of binocular stereo vision, relates to the problem of obtaining a three-dimensional face model based on stereo matching, and in particular relates to an algorithm combining area growth and dynamic programming, combined with face feature information, to obtain three-dimensional face points approach to the cloud model. Background technique: [0002] Stereo matching technology is a passive three-dimensional measurement technology, which restores the three-dimensional information of the measured object by establishing the correspondence between the pixels of two images of the measured object taken from different angles. Stereo matching technology can obtain dense 3D point cloud data of the measured object without projecting a grating to the measured object, has a high degree of automation, and can realize real-time acquisition. In the field of 3D face reconstruction, stereo matching is a commonly used techniqu...

AI Technical Summary

Problems solved by technology

This one-dimensional smooth constraint effect is limited, and it is prone to misalignment, wire drawing, etc.

[0009] (2) Lack of direction control

However, the consideration of this factor is missing in these two algorithms

[0010] (3) There is error propagation in both algorithms

But neither of these two algorithms can use these auxiliary information, resulting in a waste of information

[0012] Due to the above shortcomings, the actual application of the region growing algorithm and the dynamic programming algorithm in the reconstruction of the 3D face model is not satisfactory.

Images