Abstract:
There has been much research about using curvelet transform for retinal analysis. Other multiscale geometrical tools, such as contourlet and wavelet, have also been used for retinal analysis, thus it is interesting to know which method performs best, especially under illumination and contrast changes. In this paper, we make a systematic comparison of curvelet and contourlet for retinal image analysis, and find the best subband irrelevant to contrast and illumination changes. © 2012 IEEE.