An Improved Method for Human Eye State Detection Based on Least Square Ellipse Fitting Algorithm

摘要

In order to accurately judge eyes open or closed state of the online learners, we propose an improved method of judging eye state based on the least square ellipse fitting algorithm under natural light. Firstly, this method models human eyes, then obtains contour maps of the eye region from the image captured by the camera. Secondly, the improved least square ellipse fitting algorithm is used to search for the optimal solution to obtain the fitting ellipses of human eyes contours. Thirdly, after accurately fitting to the human eyes, the fitting ellipses are adjusted according to the threshold to obtain the actual width and height of human eyes. Finally, the human eyes state is judged by the actual aspect ratio. The experimental result shows that the accuracy of human eyes state detection based on this improved method is 91.9% (without glasses) and 85.0% (with glasses). Meanwhile, this improved method can eliminate the interference of glasses well, and compared with other traditional methods, it has higher accuracy and robustness, and has certain practical value.