Semantic Kernel Updating for Content-Based Image Retrieval
Abstract
A lot of relevance feedback methods have been pro- posed to deal with Content-Based Image Retrieval (CBIR) problems. Their goal is to interactively learn the seman- tic queries that users have in mind. Interaction is used to fill the gap between the semantic meaning and the low-level im- age representations. The purpose of this article is to analyze how to merge all the semantic information that users provided to the system during past retrieval sessions. We propose an approach to exploit the knowledge provided by user interaction based on binary annotations (relevant or irrelevant images). Such se- mantic annotations may be integrated in the similarity ma- trix of the database images. This similarity matrix is ana- lyzed in the kernel matrix framework. In this context, a ker- nel adaptation method is proposed, but taking care of pre- serving the properties of kernels. Using this approach, a se- mantic kernel is incrementally learnt. To deal with practical constraint implementations, an eigendecomposition of the whole matrix is considered, and a efficient scheme is proposed to compute a low-rank ap- proximated kernel matrix. It allows a strict control of the required memory space and of the algorithm complexity, which is linear to the database size. Experiments have been carried out on a large generalist database in order to validate the approach.
Domains
Machine Learning [stat.ML]
Origin : Publisher files allowed on an open archive
Loading...