Khosravi, J., Shams Esfand Abadi, Mohammad, Ebrahimpour, R.. (1397). Image Registration Based on Sum of Square Difference Cost Function. فناوری آموزش, 6(2), 273-281. doi: 10.22061/jecei.2019.5544.235
J. Khosravi; Mohammad Shams Esfand Abadi; R. Ebrahimpour. "Image Registration Based on Sum of Square Difference Cost Function". فناوری آموزش, 6, 2, 1397, 273-281. doi: 10.22061/jecei.2019.5544.235
Khosravi, J., Shams Esfand Abadi, Mohammad, Ebrahimpour, R.. (1397). 'Image Registration Based on Sum of Square Difference Cost Function', فناوری آموزش, 6(2), pp. 273-281. doi: 10.22061/jecei.2019.5544.235
Khosravi, J., Shams Esfand Abadi, Mohammad, Ebrahimpour, R.. Image Registration Based on Sum of Square Difference Cost Function. فناوری آموزش, 1397; 6(2): 273-281. doi: 10.22061/jecei.2019.5544.235
1Faculty of Electrical Engineering, Shahid Rajaee Teacher Training University, P.O.Box:16785-163, Tehran, Iran.
2Electrical Engineering Department, Shahid Rajaee Teacher Training University, Tehran, Iran
3SRTTU
تاریخ دریافت: 02 مهر 1396،
تاریخ بازنگری: 18 دی 1396،
تاریخ پذیرش: 25 اردیبهشت 1397
چکیده
Background and Objectives: There are numerous applications for image registration (IR). The main purpose of the IR is to find a map between two different situation images. In this way, the main objective is to find this map to reconstruct the target image as optimum as possible. Methods: Needless to say, the IR task is an optimization problem. As the optimization method, although the evolutionary ones are sometimes more effective in escaping the local minima, their speed is not emulated the mathematical ones at all. In this paper, we employed a mathematical framework based on the Newton method. This framework is suitable for any efficient cost function. Yet we used the sum of square difference (SSD). We also provided an effective strategy in order to avoid sticking in the local minima. Results: The proposed newton method with SSD as a cost function expresses more decent speed and accuracy in comparison to Gradient descent and genetic algorithms methods based on presented criteria. By considering SSD as the model cost function, the proposed method is able to introduce, respectively, accurate and fast registration method which could be exploited by the relevant applications. Simulation results indicate the effectiveness of the proposed model. Conclusion: The proposed innovative method based on the Newton optimization technique on separate cost functions is able to outperform regular Gradient descent and genetic algorithms. The presented framework is not based on any specific cost function, so any innovative cost functions could be effectively employed by our approach. Whether the objective is to reach accurate or fast results, the proposed method could be investigated accordingly.