A Novel Method For Image Retrieval Based On Visually Significant Feature Point Maps

Abstract

In this paper, we present a new algorithm for visually significant feature points detection for efficient image indexing and retrieval. We propose a hybrid feature point detector based on edge maps obtained via spatial and wavelet methods. Proposed method is compared with other existing interest point detectors. Robustness of feature points to geometric transformations like rotation and scaling is analyzed. Hausdorff distance measure is used for indexing and ranking of results.

Keywords

• Wavelets
• Edge maps
• feature points
• Hausdorff distance
• Retrieval