Efficient Content-based Image Retrieval Using Weighted Feature Aggregation Scheme

Abstract

This paper presents a content-based image retrieval system for aggregation and combination of different image features. Feature aggregation is important technique in general content-based image retrieval systems that employ multiple visual features to characterize image content. We introduced and evaluated linear combination to fuse different features. The most important step in the feature aggregation is to find suitable weights for the individual features. We have used relevance feedback techniques to determine the salient features and to learn weights for each feature. The weights are used in linear combination scheme that we call weighted feature aggregation. The implemented system has several advantages over the existing content-based image retrieval systems. Several implemented features included in our system allow the user to adapt the system to the query image. The weighted combination of features allows flexible query formulations and helps processing specific queries for which users have no knowledge about any suitable descriptors.