In this paper an automatic unsupervised method for the segmentation of retinal vessels is proposed. A self-organizing map is trained on a portion of the same image that is tested and K-means clustering algorithm is used to divide the map units in2 classes. The entire image is again input for the self-organizing map, and the class ofeach pixel will be the class of the best matching unit on the self-organizing map. Finally, the vessel network is post-processed using a hill climbing strategy on the connected components of the segmented image.The experimental evaluation on the publicly available DRIVE database shows accurateextraction of vessels network and a good agreement between our segmentation andthe ground truth. The mean accuracy is 0.9459 with a standard deviation of 0.0094 isoutperforming the manual segmentation rates obtained by other widely used unsupervisedmethods. A good kappa value of 0.6562 is inline with state-of-the-art supervisedand unsupervised approaches.
|Title of host publication||Computational Intelligence Methods for Bioinformatics and Biostatistics, 7th International meeting, Cibb 2010, Palermo, Italy, September 2010, Revised Selected Papers|
|Number of pages||12|
|Publication status||Published - 2011|
|Name||LECTURE NOTES IN COMPUTER SCIENCE|
- Theoretical Computer Science
- General Computer Science