Image retrieval algorithm based on object region location and feature fusion
Aijun Sang · Journal of Jilin University · 2008
A novel image retrieval algorithm based on local feature points is presented.First,a color image is converted into a grey image,on which the Harris operator is employed to detect corner points,and object region is located according to the distribution information of corner points.Then, the color feature and spatial feature are extracted from the color object region to represent the content of an image.The experiment results show that this algorithm is superior to those algorithms based on global feature with respect to the recall rate and precision rate.