Image Localization Combined with SOM for Color-Based Image Retrieval

A new technique to improve the performance of color based image retrieval is proposed. The proposed technique consists of two major phases. In the first phase, the SOM - self organizing map - is used to cluster images database based on some primitive features of the image. The result obtained through the first phase is used as a filtered set of images to be fed into the second phase which relies on the feature extracted and compared based on image localization to preserve the spatial similarity between search image and images database. The proposed approach overcomes the computational complexity of applying bin-to-bin comparison as a multi dimensional feature vectors in the original color histogram approach