Fatigue Recognition And Analysis Based On Gabor Feature

With the accelerating pace of modern life, fatigue phenomenon more frequent.Fatigue not only the individual’s own negative impact on mental and physical, andsometimes therefore threaten lives and safety of others. Therefore, for fatigue researchwork has been increasingly recognized by the whole community. Identified in theanalysis of fatigue and abroad on the basis of face detection, this paper presents asequence based on continuous fatigue facial expression recognition method.Diversity against human face images, this paper presents a fusion of skindetection and Adaboost face detection methods. First, determine whether a colorimage, if the color image is fused and a color space based on the color imagemorphology characteristic rough detection, to obtain a color image of the face regionextracted after Haar-Like features Adaboost classifier training and testing, the afterdetecting the face region geometric and grayscale images standardization. Reuse faceunique biological characteristics and established human face model Gabor featureextraction and data normalization, followed by LLE dimensionality reduction andsmoothing in order to better analyze and identify the fatigue characteristics. Followedby analysis of image sequences characteristic curve trends, understand the fatigueexpression changes in the time domain, on the basis of a set of proposed expressionimages as a continuous image sequence feature fatigue recognition input method andthe use of SVM identification was verified.The main work is not only enrich existing fatigue recognition and detectionalgorithms, and argue against a group of consecutive images fatigue recognition isvalid. Experiments show that the face image sequences based on continuous fatiguecharacteristics of the model can quickly solve the problem of fatigue recognition andexpression on fatigue characteristic sequence can be real and emotional fatigueintervention to provide a compensation research ideas.