Approximate nearest neighbor search using self-organizing map clustering for face recognition system

This paper presents face recognition system that is based on Self-Organizing Map (SOM) clustering. In order to reduce the time consumption in nearest neighbor search, SOM clustering scheme is used to group the training data and determine prototypes of each group. Local feature selection process is employed to reduce dimension of data in each group. To show the performance of the proposed scheme over various choices of feature extraction method, PCA (Eigenface), 2DPCA, and SOM-Face are tested in the experiment. Recognition accuracy and time consumption are measured in comparison with k-d Tree search and the other clustering based search schemes by using the dataset of 1,560 face images from 156 people. The experiments show that the proposed scheme can obtain the best recognition rate of 99.36% while it reduces the time consumption.