Study Of Mental Fatigue Recognition Based On Static Facial Features

The rapid development of science and technology,physical labor in daily life has been gradually replaced by mental work,therefore,the probability of mental fatigue will be greatly increased and the impact of it will also increase,almost all areas are involved,especially in the driver who is more significant.Fatigue recognition based on facial and facial images,first using a single feature of the facial features fatigue identification,and then use a variety of feature fusion method to identify fatigue,although some valuable achievements have been made on the basis of the existing ones,the times are progressing and more and more problems will arise.Therefore,paper presents a study of mental fatigue recognition based on static facial visual features.Based on the image processing and fatigue identification technology,this thesis studies the face mental fatigue recognition algorithm based on facial visual features and fatigue recognition,and uses Matlab as the simulation tool to realize the facial recognition fatigue identification.The main contents include face image preprocessing,face rough positioning and face accurate positioning,face visual features eyes and mouth coarse positioning and precise positioning,single feature of the eyes or mouth mental fatigue identification,fusion of multiple features of the spirit Fatigue recognition.The concrete realization process is,first,using the integral projection algorithm to coarsely locate the eye features,and then use the improved pupil detection algorithm to locate the eye area in the coarse positioning area.The core idea of the improved pupil detection algorithm is to search the coarse positioning area The largest number of black pixels,with its pupil center draw a circle,and finally use the least squares fit pupil circle.In the same way,firstly,the integral projection algorithm is used to coarsely position the mouth features,and then the improved Harris corner detection algorithm is used to accurately position the coarse positioning mouth.The core idea of the improved Harris corner detection algorithm is to eliminate some pseudo-corners first,Determine the corner point candidates,and then use the adaptive threshold to determine the final corner point.Finally,the improved SVM model is used to separately identify the fatigue characteristics of the single visual features and the merged features.The improved support vector machine is a local AdaBoost algorithm.Through Matlab simulation test,the results show that the algorithm of this study is intuitive and effective,and the ability of fatigue recognition has been improved.