An Efficient Content Based Image Retrieval System Using Soft Computing Technique

As the use of searching digital images has increased a lot in recent years, so a system is required which search images from large database on the bases of query image and return results in form of images which matched with query image. The present paper introduces a content based image retrieval system using artificial neural network (ANN) approach as soft computing technique. The proposed system is composed of three major phases: feature extraction, ANN classifier and matching strategy. In feature extraction phase, local and global features are extracted and then Gabor filter is applied to enhance images so that better results can be calculated. The artificial neural network (ANN) in our proposed system serves as a classifier so that selected features of a query image are used as an input to find multi classes as an output which have largest similarity to the query image. At last feature matching strategy is used to retrieve images from database as result. Finally the results presented in this paper by using proposed system are accurate and takes less computation time. Keywords— ANN, CBIR, Color and Texture features, Similarity measures, Classification