Object Recognition Using Modified Normalized Cross Correlation

Normalized cross correlation is one of the practical methods for comparing the similarity of the two images.This paper presents a new method to detect objects in the picture based on normalized cross correlation. In this method,the reference image is segmented using color segmentation based on K-means clustering method and various objects in the reference image are extracted.Then, the similarity between pattern binary image and objects binary images, extracted from reference image,is examined using normalized correlation. If rotated desired patterns exist in the reference image, it will be recognized with this method.