Content-based Semantic Indexing of Image Using Fuzzy Support Vector Machines

With the increasing amount of multimedia data, content-based image retrieval attracts many researchers of various fields in an effort to automate data analysis and indexing. In this paper, we propose a content-based semantic indexing method which annotates images automatically using concepts and textual description. In order to bridge the "semantic gap" between the low-level descriptors and the high-level semantic concepts of an image, we introduce a 3-level pyramid and combine the color, texture and edge features for each level. Fuzzy Support Vector Machine (FSVM) is employed for building the concept model and calculates the likelihood of an image to a model. We index the images with concepts according to the likelihood between an image and the concept model. Experiments show that our method has good accuracy in semantic indexing of images.