A fast image texture analysis algorithm
Zhang Tao · Guangdian gongcheng · 2005
In view of the faults of complex calculation and long operation time existed in the present texture analysis algorithms, a new image texture analysis algorithm is proposed. The difference matrix and difference histogram of an image is defined by difference concept, the likelihood measurement for image texture is carried out through extracting features from difference histogram. Since the adding and subtracting operation is adopted as much as possible on the basis of taking account of human vision behavior, the algorithm not only has the features relating to human vision behavior, its operation efficiency also has been greatly improved. The retrieval experiments for brodatz image database are carried out by this algorithm. The results show that compared with the Gray Level Co-Matrix (GLCM) algorithm the accuracy and retrieval efficiency of the algorithm have been improved for 13.9160% and 9.5 times, respectively. It has a better retrieval effect and higher retrieval efficiency than that of existing algorithms.