A Human Eye Positioning Method Based on Nose-Eye Structural Constraints

Abstract

The invention relates to a human eye positioning method based on the constraints of the nose and eye structure, which belongs to the field of human eye positioning technology, and involves four steps: face detection, human eye and nostril detection, filtering out error points, and missing detection based on template matching. Real-time detection of the position of the human eye is completed. Compared with the prior art, the present invention outputs a group of nose and eye feature point data clusters with the highest reliability based on a mixed probability model of a large number of nose and eye feature point data, and filters out unreasonable outputs. To detect technical problems that are often disturbed and cannot be accurately detected, introduce nostril detection that is not easily disturbed and has high detection accuracy, and perform all the detected nose and eye feature point data according to the structural correlation of the nose and eye distribution in the face. Data clustering based on the mixed probability model can improve the probability of correct human eye position and increase its credibility.

Description

technical field [0001] The invention belongs to the technical field of human eye positioning, and relates to a human eye positioning method based on nose-eye structural constraints. Background technique [0002] With the continuous development of society, human eye positioning technology is widely used in monitoring, criminal investigation, transportation and medical treatment. Its technology combines aspects of biology, image processing, computer vision, and pattern recognition. With the continuous development of intelligent interaction technology, technologies closely related to human eye positioning technology have emerged, such as face recognition, gaze tracking, virtual reality and augmented reality. However, the commonly used human eye positioning method is based on the grayscale of the human eye itself. degree distribution to locate human eyes. For example, the existing gray-scale integral histogram method uses the trough distribution of human eyes on the gray-scale ...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the purpose of the present invention is to provide an image feature algorithm combined with nose-eye structural constraints for the existing technical problems that are easily affected by their respective feature defects and prone to multiple false detections that affect the detection accuracy of the human eye. The positioning method and system assists the positioning of human eyes by detecting nostrils with high accuracy, reduces the false detection caused by a single image feature detection algorithm, and combines template matching to speed up detection. The system can effectively improve the detection accuracy of human eyes And the real-time performance of human eye detection

Images