Face Recognition Based On Statistical Feature And Illumination Compensation

Face recognition is one of the most important research fields of image processing,pattern recognition and artificial intelligence, its purpose is to identify the identity of thepeople using computer through face feature. It has a wide range of applications, includingcommerce, security, person verification, and law enforcement. Face recognition based onstatistical characteristics is now one of the most concerned face recognition technology bythe researchers, it overcomes the shortcomings appeared in the process of traditional facerecognition methods, using statistical knowledge can be effective for face recognition.However, since face patterns are complicated and multiform, the face recognition rate willface sharp decline under various conditions, such as changing illumination, pose and facialexpression, which are very difficult to represent face effectively. Therefore, face recognitionmethods with robustness and efficiency are hotspots in recent studies.The thesis studies on feature extraction problem in face recognition, and the focusesare the combination of global feature and local feature. Our goal is to further improve therobustness with varying expression, illumination and shadow while improving the facerecognition rate at the same time. The main work and innovations are as follows:1. This paper analyzes LBP algorithms, then combines it with PCA algorithm, presentsa (2D)~2PCA-LBP algorithm of face recognition. This method first extracts facial texturefeature, then uses (2D)~2PCA algorithm to reduce its dimension. The reason is that LBPalgorithm has a characteristic of rotation invariance, which is robustness to illuminationchanging and pose variation.(2D)~2PCA algorithm is the improvement of PCA, using thismethod, the image can reach the maximum degree of dimensionality reduction. Theexperiment results show that the algorithm can improve the face recognition rate, especiallyto the image with illumination changing, pose variation and facial expression, therecognition rate improve significantly.2. The thesis studies the compressed sensing method in-depth. In order to solve therobustness problem with block, expression and illumination in face recognition system, wepropose a face recognition method based on PCA-based compressed sensing algorithm.Utilizing (2D)~2PCA transform to extract image features in both row and column directionsand reducing the dimension. A projection matrix is constructed to identify the face features,considering these features to form an over complete dictionary. By solving the l1norm minimization, seeking out the sparsest representation of images based on the dictionary toobtain a set of optimal sparse coefficients, which are used to recover the train images,compute the residuals between test and train images for face recognition. The method breaksthrough the characteristics of traditional method using only one class for recognition, we useall the training set for classification. The classifications results are improved and the timecomplexity is reduced to linear order, in the same time, the recognition rate improvessignificantly. In addition, as long as the training samples are sufficient enough, the testsamples can be effectively expressed. Especially for huge changes in pose variation,illumination and expression, the recognition rate has been significantly improved.