![]() | Only 14 pages are availabe for public view |
Abstract Nonextensive entropy image thresholding is a powerful technique for image segmentation. The presented method [El-SayedA, 2011] has been derived from the generalized entropy concepts proposed by Tsallis of order q. The advantage of the method is the use of a global and objective property of the two-dimensional histogram and this method is easily implemented. In almost every image used, the proposed method yielded a good threshold value for fractional q coefficient, that is, when 0<q<1. For q>1, the proposed method did not yield a good threshold value. When q = 1, the method yielded, in some cases, good optimal threshold values and, in some other cases, unacceptable threshold values. |