Image retrieval using NN based pre-classification and fuzzy relevance feedback

In this article, we have proposed an interactive image retrieval scheme using MPEG-7 visual features, Neural Network (NN)-based pre-classifier and fuzzy based feature evaluation scheme. The performance of the existing image retrieval systems is generally limited due to semantic gap, resulted due to the discrepancies between the computed low-level features and user's conception of an image. Partitioning the database by a NN-based pre-classifier, and using a fuzzy based feature evaluation scheme, the performance of the proposed scheme has been found to improved drastically by reducing the retrieval time and increasing the accuracy.