|
In this paper we present a new method for visual word based similar image retrieval by comparing content of a query image with images stored in a database. The retrieval consists of three main steps: feature extraction, indexing and query optimization. The feature extraction step is based on SURF algorithm. For indexing, we use the K-Means algorithm and the Bag-of-Visual-Words model. The last step is very significant and we associate TF-IDF with Hamming Distance to query. Our method is tested on the highly diverse opening images and has proved a better retrieval accuracy based on the experimental results. |
|
Keywords:Content-Based Image Retrieval (CBIR), Speed-Up Robust Features (SURF), Bag-of-Visual-Words (BoVW), K-Means, Hamming Distance Code |
|