Knowledge-based image retrieval system

作者:

Highlights:

摘要

Most of the retrieval systems concentrate much on low-level features such as color, texture, shape and position. The present system is mainly developed based on the visual descriptors of the image such as color, texture and shape descriptors, etc. along with the high-level semantic analysis of the image content through different processing modules in the proposed architecture. Similarity measures are proposed and the performance evaluation has been done. As an image browser, apart from retrieving images by image example, it also supports query by natural language. The present system works well both online and offline.We have used unsupervised Kohonen’s self-organizing maps (SOM) technique to train the images and our own indexing scheme with reference system based on R-tree SOM. We proposed an approach fuzzy color histogram for color retrieval, and Lie descriptors for the retrieval of shapes. We have also tested our appraoch for MPEG-7 images.

论文关键词:Relevance feedback,Fuzzy color histogram,Baysesian distance,Lie algebra,Diffeomorphisms,Semantic analysis,Knowledge base

论文评审过程:Received 1 March 2006, Accepted 1 February 2007, Available online 12 February 2007.

论文官网地址:https://doi.org/10.1016/j.knosys.2007.02.002