A User Interface for Query-by-Sketch based Image Retrieval with Color Sketches

Ivan Giangreco, Michael Springmann, Ihab Al Kabary, Heiko Schuldt
In Proceedings
Appears in
Proceedings of the European Conference on Information Retrieval (ECIR)
Barcelona, Spain
Springer Berlin / Heidelberg

Over the last few years, the sizes of image archives have been rapidly growing which makes querying for specific images, also referred to as known item search, increasingly difficult. As more and more images are not annotated with proper metadata, content-based image retrieval (CBIR) is a viable approach to image retrieval that solely relies on the inherent visual characteristics. A major problem in CBIR is however the availability of a query image that is good enough to express the user's information need. Query-by-sketch allows  to overcome this limitations as users can contribute sketches as query input. In contrast to most existing approaches that consider only edge information extracted from the sketch, this demo presents a system that jointly exploits edge and color information, gathered via a novel, user-friendly interface.