An Improved Method for Face Recognition Based on Sparse Representation

In this paper, a face recognition method based on sparse representation is proposed. First it represents the test sample as a linear combination of all the training samples and selects K (K = 1, 2, 3) training samples from every class with the bigger contribution by using the representation result of every test sample. Then the test samples are represented as a linear combination of the selected training samples and we determine the classification of the test samples by distance formula method. Besides, some experiments are carried out with this algorithm in different distance formulas, which shows that Euclidean distance method is the best. The experimental results obtaining from different face database demonstrate the validity of our theoretical prediction.