This is very interesting and helpfull........[/quote]
wow nice...thank you.. this is going to be a very helpfull topic for me thanks for uploading this ....
I have need some matlab coding for optic cup image segmentation and going to be find out how to helpfull for me
thanks
thank you very much its very helpful oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
Posts: 14,118
Threads: 61
Joined: Oct 2014
A circular transformation is designed to simultaneously capture the circular shape of the DO and the variation of the image across the OD boundary. For each retinal image pixel, it evaluates the variation of the image along multiple uniformly oriented radial line segments of specific length. Pixels with maximum variation along all segments of the radial line are determined, which can be further exploited to accurately locate the OD center and the OD boundary. Experiments show that the 99.7%, 97.5% and 98.77% OD detection accuracies are obtained for the STARE data set, the ARIA data set and the MESSIDOR data set, respectively, and the error of the OD center is about six pixels for the data set STARE and ARIA which is much smaller than that of cutting-edge methods ranging from 14-29 pixels. In addition, the segmentation accuracy of 93.4% and 91.7% for the STARE dataset and ARIA dataset, respectively, are obtained, consisting of many severely degraded pathological retinas that the methods of can not segment correctly. In addition, the algorithm runs in 5 s, which is substantially faster than many of the methods of the prior art.