Fast eyelid detection method and system based on gabor filter

Abstract

The invention discloses a fast eyelid detection method and system based on gabor filtering. The method comprises the following steps: obtaining an eye picture, dividing the upper eyelid area and the lower eyelid area in the eye picture; The regions are filtered separately to obtain the upper eyelid gradient map and the lower eyelid gradient map; based on the pixel point with the maximum energy value in each column of pixels in the upper eyelid gradient map and the lower eyelid gradient map, the upper eyelid candidate point and the lower eyelid candidate point are respectively obtained; Based on the upper eyelid candidate point, obtain the upper eyelid candidate line segment, and obtain the upper eyelid effective point according to the upper eyelid candidate line segment; based on the lower eyelid candidate point, obtain the lower eyelid candidate line segment, and obtain the lower eyelid effective point according to the lower eyelid candidate line segment; The quadratic polynomial fitting is performed on the effective points of the upper eyelid and the effective points of the lower eyelid respectively, and the fitting curve of the upper eyelid and the fitting curve of the lower eyelid are obtained. The eyelid detection method of the present invention is not only efficient, but also has strong robustness.

Description

technical field [0001] The invention relates to the technical field of biological identification, in particular to a fast eyelid detection method and system based on gabor filtering. Background technique [0002] Iris recognition is the most accurate biometric technology and is widely considered to be the most promising biometric technology in the 21st century. Future applications in various fields such as security, national defense, and e-commerce will inevitably focus on iris recognition technology. This trend has begun to appear in various applications around the world, and the market application prospect is very broad. [0003] However, in the process of collecting iris images, the tester may blink and so on. How to judge whether the available iris area in the collected iris images meets the requirements of feature extraction has become an important task in iris recognition technology. [0004] The existing eyelid detection methods mainly include the following: [000...

AI Technical Summary

Problems solved by technology

[0008] However, the above algorithm takes a long time to detect. For example, Adam's algorithm needs 300 milliseconds to detect an eye diagram, which has low detection efficiency and poor robustness.

Images