Multi-object face recognition using Content Based Image Retrieval (CBIR)

Abstract

Real-time face recognition system process divided into three steps, feature extraction, clustering, detection, and recognition. Each step uses a different method that is Local Binary Pattern (LBP), Agglomerative Hierarchical Clustering (AHC) and Euclidean Distance. Content Based Image Retrieval (CBIR), an image searching techniques based on image feature, is implemented as the searching method. Based experiments and the testing result, recall and precision values are 65.32% and 64.93% respectively.