An Adaptive Approach towards Content-based Image Retrieval

We propose and evaluate an adaptive approach towards content-based image retrieval (CBIR), which is based on the Ostensive Model of developing information needs. We use ostensive relevance to capture the user’s current need and tailor the retrieval accordingly. Our approach supports content-assisted browsing, by incorporating an adaptive query learning scheme based on implicit user feedback. Textual and colour features are employed to characterise images, which are combined using the Dempster-Shafer theory of evidence combination. Results from a user-centred, task-oriented evaluation show that the ostensive interface is preferred over a traditional interface with manual query facilities. Its strengths lie in its ability to adapt to the user’s need, and its very intuitive and fluid way of operation.