Year: 2006
Type: article
Abstract: . We present a method for image retrieval by color and shape similarity matching using metric indexing. We extract color features into the HSV color and for shape features extraction we use Fourier descriptors. We index the combined color and shape features vectors into a M-tree. We complement the use of the M-tree with a filtering technique that also permits partial-image matching. We use a similarity measure calculated as an affine linear combination of L1 and L2 metrics. In the refinement query step for color extraction we use a special histogram with a uniform color transition which enables a window-based smoothing during retrieval. Our experimental results show that our technique is both correct and efficient.
Author: Maria M. Ruxanda
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W1782310529
