Using Patch Histogram of Detected Corners for Image Retrieval

In this paper, we proposed an algorithm which extracts patches from corner points for calculating their histogram. Then that histogram is used to construct a feature vector for representing an image and subsequently for searching the similar images from image database. The corners are detected and extracted by finding the intersections of the detected lines. Lines are identified by using Hough transform. Before detecting the lines we have to find out the edges information from the image, this is done by using canny edge detection algorithm. Taking the corner as the center coordinate, the 9 × 9 pixel patch is extracted to get the pixel information around the corner is used to calculate the histogram. This result in a significant small size feature matrix compared to the algorithms using color features. Experimental results show that it is computationally efficient and the image corners being noise invariant produce good results in noisy environments.