Different relevance feedback techniques in CBIR: A survey and comparative study

In Content Based Image Retrieval (CBIR), there is semantic gap between the low level features and high level concepts. Advent of different relevance feedback techniques bridges the gap. Different techniques use different assumption. Yet there are some performance metrics which quantitatively measure and compare different relevance feedback algorithms. This is necessary to make the CBIR system work consistently. In this paper we analyze different relevance feedback algorithm starting from conventional to recent technique.