Mathematical snalysis of blood vessels segmentation based on three retinal images datasets
Retinal images are routinely acquired and retinal blood vessels are segmented to provide diagnostic evidence of diabetic retinopathy. Because of the acquisition process, very often these images are non-uniformly illuminated and exhibit local luminosity and contrast variability. In this work, three b...
| Main Authors: | , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
2014
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/38434/ http://irep.iium.edu.my/38434/ http://irep.iium.edu.my/38434/1/Eye_Vessel.pdf |
| Summary: | Retinal images are routinely acquired and retinal blood vessels are segmented to provide diagnostic evidence of diabetic retinopathy. Because of the acquisition process, very often these images are non-uniformly illuminated and exhibit local luminosity and contrast variability. In this work, three basic algorithms for segmenting retinal blood vessels, based on different image processing techniques which are Matched filter, Hough transform and Morphological operations are used and their strengths and weaknesses in terms of retinal images segmentation are compared mathematically. The performance of each algorithm was tested on DRIVE, DRIONS and High-Resolution Fundus images database. The results show that measuring the automatic segmentation algorithm performance is based mainly on how the retinal images are acquired as well as the image processing technique used for segmentation. The obtained
results could help the eye specialists in their visual examination of retinal images. |
|---|