Face Image Recognition Algorithm based on Singular Value Decomposition

Abstract

In face image recognition, features play a decisive role in recognition and classification. Feature extraction can describe the image, but the extracted data may contain redundant and useless information, which affects the model generalization learning. For these problems, a face image recognition algorithm with singular value decomposition is proposed. The original data is firstly decomposed by singular value decomposition with SVD algorithm, and then the k values in the top of the singular value are selected and calculated to obtain the sample information after attribute reduction, and then the result of the reduced data is classified by the model using the extreme learning machine algorithm, and finally the type corresponding to the image can be predicted by the model, and the experimental results on the ORL face image data set also prove that algorithm has a good recognition efficiency.