Study on Image Retrieval based on Image Texture and Color Statistical Projection
Description
This article presents an image texture and hue statistical projection based retrieval. First the image is converted to HSI color model, the gray value of the image extraction, and Robert algorithm to extract the texture, then the image is divided into blocks and extracts the main color block, the main color image blocks are respectively projected in the horizontal and vertical direction of 2, get 2 projection histogram, the 2 projection histograms of the first three order center extraction distance and Robert algorithm as the features of texture, image similarity calculation. Make a very full pave the way for future Canny edge processing algorithm research of image retrieval.
Files
Steps to reproduce
Firstly, the image is converted to HSI color model, the gray value of the image extraction, and Robert algorithm to extract the texture; then, the image is divided into blocks and extracts the main color block, the main color image blocks are respectively projected in the horizontal and vertical direction of 2, get 2 projection histogram, the 2 projection histograms of the first three order center extraction distance and Robert algorithm as the features of texture image similarity calculation; finally,make a very full pave the way for future Canny edge processing algorithm research of image retrieval.