Using LIRE for spase visual word-based image retrieval
LIRE, which stands for Lucene Image REtrieval, is a lucene-based image content indexing and search engine developed by Mathias Lux.
This project uses LIRE to support sparse visual word-based CBIR (content-based image retrieval). It modifies partly the codes of LIRE. The codes for this project are found at lire-sparsevw/. There are codes for LIRE indexing and search engine, and web interface of PHP and HTML.
The binary, extract_features_64bit.ln, in web part is interest point detectors/descriptors implemented by Krystian.Mikolajczyk@inrialpes.fr at INRIA Rhone-Alpes.[ref. www.inrialpes.fr/movi/people/Mikolajczyk/Affine].