Image based human canthus detection method and system

Abstract

An image based human eye corner detection method and system in the field of image processing are disclosed. A rough eye region image and an eye corner region image are filtered by using a fixed template and are binarized with a threshold self-defining method; a similar eye corner point image and an eye white image in the rough eye region image are extracted; coordinate information of a most right point in the similar eye corner point image is utilized as a boundary; a communicated region corresponding to the right side of the boundary in a rough eye region binary image obtained in the second step is filtered away; a primary corrected rough eye region binary image is obtained; the primary corrected rough eye region binary image and the eye white image are combined; a secondary corrected rough eye region binary image is obtained; the relations between two largest communicated regions in an eye corner region binary image and the rough eye region binary image are calculated so as to obtain corresponding communicated regions of the two communicated regions in the rough eye region binary image; the relations between the two communicated regions and the communicated region of the secondary corrected rough eye region binary image are calculated; whether a rough eye region and an eye corner region need to be re-positioned is judged; and finally the eye corner region binary image is pruned and eye corner point information is output. With the adoption of the detection method and system, the eye corner detection can be effectively realized.

Description

technical field [0001] The present invention relates to a technology in the field of image processing, in particular to an image-based method and system for detecting corners of human eyes. Background technique [0002] Face key point technology is to locate the key features of the face, such as eyes, eye corners, nose, mouth, etc. These key points are rich in a lot of information, which can provide corresponding basic data for research work such as face recognition, expression analysis, and face tracking. The corner of the eye as a key point is involved in various aspects, such as the Gaze application for estimating the human eye line of sight, or the information extracted from the eye area in the judgment of the human eye state. [0003] After searching the literature of the prior art, it was found that P.N.Belhumeur et al published an article called Localizing parts of faces using a consensus of exemplars (using the consistency of examples to determine the part of the fa...

AI Technical Summary

Problems solved by technology

[0005] Although the above methods can detect the position of the corner of the eye, in face key point detection, because the position of other key points on the face needs to be located, the individual information of the corner of the eye is often limited by other parts, and sometimes even by other key points. However, there are very few existing methods for directly detecting the position of the corner of the eye, and the only methods are usually unable to cope with changes in light, different biological structures of the human eye, etc.

In some special cases, the movement of the iris of the human eye will also make the detection inaccurate

Images