Image retrieval based on combining local dominant color and edge histogram
Tai De-yan · Journal of Guangxi University of Technology · 2009
The use of color histogram is now a classical method in the CBIR,which has proved to be limited.An image retrieval method based on local primary color-edge multi-features is proposed.In the field of retrievaling based on color,we use HSV color model to quantify the color and design 25 overlapped sub-regions to get color histogram in each region.Image edge is extracted by using Sobel detection operator.2D edge histogram is used to describe shape feature.By assigning a weighting value,the technology based on color and edge features combine to a multimethod.The result shows that the retrieval results obtained from combined-features are superior to that obtained from single-features.