Blood Vessels Segmentation in Retina image using Mathematical Morphology and the STFT analysis

This paper presents an algorithm based on mathematical morphology and short time Fourier transform (STFT) analysis for the detection of the blood vessels in retinal images. In a first step, blood vessels are enhanced by mathematical morphology filters such as "supremum of opening", "toggle mapping" transformation and "morphology reconstruction". In a second step, the result is refined by an STFT analysis and a contextual filtering in the Fourier domain. The method is also able to yield the local ridge orientation and ridge frequency information, and it\´s able to successfully enhance and reconstruct the blood vessels. Nevertheless, it is worth mentioning that the mathematical morphology method was able to detect finer detail more precisely. Our present results suggest that an interesting direction to be investigated is how to use both approaches together in order to obtain better results and apply this as a diagnostic tool. Our method has been tested on the Utrecht database by M. Niemeijer et al. (2004) consisting of 40 manually labeled images. The method achieves an area Az under the receiver operating characteristic curve (ROC) of 0.961. The method is compared with three recently published rule-based methods of Chaudhuri et al. (1989) and Martinez-Perez et al. (1999) and Zana et al. (2001). The results show that our method is significantly better than the three rule-based methods