Content-based image retrieval using multiresolution speeded-up robust feature

Abstract

The advent of numerous low cost image capturing devices has led to the proliferation of huge amount of images in the present world. The images have grown more complex day-by-day and in order to access them easily, there is a need of efficient indexing and retrieval of these images. The field of content-based image retrieval (CBIR) tends to achieve this goal. This paper proposes the concept of multiresolution speeded-up robust feature (SURF) descriptor which combines discrete wavelet transform and SURF descriptor to extract interest points at multiple resolutions of image for CBIR. The feature vector has been constructed through grey-level co-occurrence matrix (GLCM). The advantage of this technique is that it exploits multiple resolutions of image to extract interest points which single resolution processing techniques fail to do. Performance of the proposed method is tested on two benchmark datasets Corel-1K and GHIM-10K and measured in terms of precision and recall. The performance of the proposed method is measured with other state-of-the-art feature descriptors. The experimental results demonstrate that the proposed method outperforms other state-of-the-art descriptors in terms of precision and recall.