Color Image Retrieval Based on a two-Dimensional Histogram
Khalid El Asnaoui, Brahim Aksasse, Mohammed Ouanan · International journal of mathematics and computation · 2015
The problem we consider that is about finding similar images in a large database. The most efficient algorithms use local image descriptors. In this paper, we propose a new algorithm based on the intersection of 2-D histograms in HSV space. The proposed histogram is based not only on the intensity of pixels but also on a 3x3 window. This approach overcomes the drawback of the classical histogram which ignores the spatial distribution of pixels in the image. By comparing its performances to several methods of the state of the art, we will show that the developed method presents several advantages. The proposed histogram is faster, reduce memory consuming and it doesn’t require learning. For validation of our results, this algorithm will be applied to search similar images in a database of over than 1000 images. Finally, we show that the proposed algorithm is efficient than the method of searching by the intersection of classical histograms in HSV and RGB spaces.