Although powerful image representations have been proposed for content-based image retrieval, most of the current systems are "rigid", i.e. they retrieve a fixed set of images as response to a given query and an image feature. In this paper, our goal is to introduce tools for making image retrieval systems more flexible. More precisely, we use multiple image features, and present in details a new relevance feedback technique that integrates the positive and negative examples provided by the user. Experimental results on various large databases show that the proposed technique is more performant than the standard relevance feedback approach.

A2thozgh pouleTful image TepT~entatiom have been proposed for cozienf-based image Tetieval, most of the cument systems aTe "tigid", i.e. they rettieve a fied set of images as Tesponse to a given que~and an image featuTe.We introduce S=fitaage, a useT-fiendly, genetic and flw-ble content-based image TetievaI system.Stiiatage wes the que~-by-sample appToach for ~ettieving images and integrate advanced ~eatuTessuch as image signatuTe combination, classification, multiple queti~and queq Refinement.The classic and advanced featuTes of SMiraage aTe detaild in the papeT.Stiiraage has been eztensive[y t~ted oz dozens of databas~and p~oauced ezcellent Tetm-evalTe-S-dts; a sample of Tetm"evaTau~ts h pr~ented here.

Relevance feedback is one of the strong components of Surfimage, the INRIA content-based image retrieval system. Relevance feedback is about learning from user interaction, and is useful in tasks like query refinement and multiple queries. We present two relevance feedback techniques currently implemented in Surfimage.