Header menu link for other important links
X
FPGA based hardware synthesis for automatic segmentation of retinal blood vessels in diabetic retinopathy images
Published in ISA - Instrumentation, Systems, and Automation Society
2014
Pages: 156 - 163
Abstract
Diabetic Retinopathy (DR) is a disorder that affects the structure of retinal blood vessels due to long-standing diabetes mellitus. Real-Time mass screening system for DR is vital for timely diagnosis and periodic screening to prevent the patient from severe visual loss. Human retinal fundus images are widely used for an automated segmentation of blood vessel and diagnosis of various blood vessel disorders. In this work, an attempt has been made to perform hardware synthesis of Kirsch template based edge detection for segmentation of blood vessels. This method is implemented using Lab VIEW software and is synthesized in field programmable gate array board to yield results in real-time application. The segmentation of blood vessels using Kirsch based edge detection is compared with other edge detection methods such as Sobel, Prewitt and Canny. The texture features such as energy, entropy, contrast, mean, homogeneity and structural feature namely ratio of vessel to vessel free area are obtained from the segmented images. The performance of segmentation is analysed in terms of sensitivity, specificity and accuracy. It is observed from the results that the Kirsch based edge detection technique segmented the edges of blood vessels better than other edge detection techniques. The ratio of vessel to vessel free area classified the normal and DR affected retinal images more significantly than other texture based features. FPGA based hardware synthesis of Kirsch edge detection method is able to differentiate normal and diseased images with high specificity (93%). This automated segmentation of retinal blood vessels system could be used in computer-assisted diagnosis for diabetic retinopathy screening in real-time application. Copyright 2014, ISA All Rights Reserved.
About the journal
Journal51st Annual Rocky Mountain Bioengineering Symposium, RMBS 2014 and 51st International ISA Biomedical Sciences Instrumentation Symposium 2014
PublisherISA - Instrumentation, Systems, and Automation Society
Open AccessNo
Concepts (15)
  •  related image
    Blood vessels
  •  related image
    Edge detection
  •  related image
    Eye protection
  •  related image
    Feature extraction
  •  related image
    Field programmable gate arrays (fpga)
  •  related image
    Hardware
  •  related image
    IMAGE TEXTURE
  •  related image
    OPHTHALMOLOGY
  •  related image
    Textures
  •  related image
    Diabetic retinopathy
  •  related image
    EDGE DETECTION METHODS
  •  related image
    FPGA HARDWARES
  •  related image
    KIRSCH TEMPLATE
  •  related image
    RETINAL FUNDUS IMAGES
  •  related image
    Image segmentation