Randomized Hough transform can effectively recognize circles in images,but it can easily introduce a large amount of invalid accumulation information,which seriously affects the accuracy and speed of pupil recognition.To address this issue,the research group has proposed an improved method.This method first randomly samples a point on the edge image,then searches for a candidate circle in the connected domain of that point,and finally uses a true or false method to determine the candidate circle for pupil recognition.This improvement method not only effectively solves the problem of randomized Hough transform easily introducing a large amount of invalid accumulated information,but also greatly improves the accuracy and speed of pupil recognition.
维普
万方数据