Affiliations: School of Computer and Information Technology, Liaoning Normal University, Dalian 116029, P.R. China. [email protected], [email protected]
Correspondence:
[†]
Address for correspondence: School of Computer and Information Technology, Liaoning Normal University, Dalian, P.R. China.
Note: [*] This work was supported by the National Natural Science Foundation of China under Grant No. 61272416.
Abstract: Moments and moment invariants have become a powerful tool in image processing owing to their image description capability and invariance property. But, conventional methods are mainly introduced to deal with the binary or gray-scale images, and the only approaches for color image always have poor color image description capability. Based on Exponent moments (EMs) and quaternion, we introduced the quaternion Exponent moments (QEMs) for describing color images in this paper, which can be seen as the generalization of EMs for gray-level images. It is shown that the QEMs can be obtained from the EMs of each color channel. We derived and analyzed the rotation, scaling, and translation (RST) invariant property of QEMs. We also discussed the problem of color image retrieval using QEMs. Experimental results are provided to illustrate the efficiency of the proposed color image descriptors.
Keywords: Color image, quaternion, Exponent moments, quaternion Exponent moments, geometric invariance, color image retrieval
