Title :
A new fusion scheme for multi-focus images based on dynamic salient weights on discriminative edge points
Author :
Chen, Cheng-I ; Lin, Phen-Lan ; Huang, Po-Whei
Author_Institution :
Dept. of Comput. Sci. & Eng., Nat. Chung Hsing Univ., Taichung, Taiwan
Abstract :
In this paper, we propose a new fusion scheme for multi-focus images based on special edge information called discriminative edge points. A new focus measure is also proposed to identify which pixels in the source images can be included in the resultant image to achieve the best fusion result by evaluating the salient weights for the discriminative edge points within a dynamic focus-measuring window whose center is the location of the pixel under current consideration. Experimental results show that our fusion scheme outperforms other seven methods reported in the literature.
Keywords :
edge detection; image fusion; discriminative edge points; dynamic salient weights; focus measurement; image sources; multifocus images; new fusion scheme; Equations; Image edge detection; Image fusion; Machine learning; Pixel; Wavelet transforms; Weight measurement; Multi-focus images; discriminative edge points; false edge problem; image fusion;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2010 International Conference on
Conference_Location :
Qingdao
Print_ISBN :
978-1-4244-6526-2
DOI :
10.1109/ICMLC.2010.5581039