Semantic real-world image classification for image retrieval with fuzzy-ART neural network

Abstract

Image retrieval based on semantic learning has attracted considerable attention in recent years. Analyzing the contents of an image and retrieving corresponding semantics are important in semantic-based image retrieval systems. Region-based image retrieval systems attempt to reduce the gap between high-level semantics and low-level features by representing images at the object level. In this paper, we apply principal component analysis to extract significant region features and then incorporate them into the proposed two-phase fuzzy adaptive resonance theory neural network (Fuzzy-ARTNN) for real-world image content classification. In general, Fuzzy-ARTNN is an unsupervised classifier. The training patterns in image content analysis are labeled with corresponding categories. This category information is useful for supervised learning. Thus, a supervised learning mechanism is added to label the category of the cluster centers derived by the Fuzzy-ARTNN. Moreover, based on the content analysis by the proposed two-phase Fuzzy-ART, each region in an image is associated with a high-level semantic concept. The proposed system supports both query by keyword(s) with/without region size and query by specified region(s). Experimental results show that the proposed method has high accuracy for semantic-based photograph content analysis and that the results of photograph content analysis are similar to the perception of human eyes. In addition, the semantic-based image retrieval system has a high retrieval rate.