Title: Diabetic retinopathy detection using local ternary pattern
Authors: A. Anitha; S. Uma Maheswari
Addresses: Department of Electronics and Communication Engineering, Government College of Technology, Coimbatore 641013, India; Department of Electronics and Communication Engineering, Coimbatore Institute of Technology, Coimbatore 641014, India ' Department of Electronics and Communication Engineering, Government College of Technology, Coimbatore 641013, India; Department of Electronics and Communication Engineering, Coimbatore Institute of Technology, Coimbatore 641014, India
Abstract: An intelligent way of diabetic retinopathy detection (DR) at an early stage is required to prevent blindness. DR is detected by analysing the retinal background without segmenting the lesions. This work focuses on local ternary pattern (LTP) for analysing texture of the fundus image. As local binary pattern (LBP) is more sensitive to noise and illumination variation, LTP is employed and its discriminative power is explored. LTP is obtained for all three colour components, red (R), green (G) and blue (B) for different radius considering eight neighbours. The histogram of LTP and variance form a feature set for the classifiers KNN and random forest with ten-fold cross validation. Random forest provides a sensitivity and specificity of 100%. The average sensitivity and specificity of nearly 91% are achieved. The proposed algorithm is very fast and can be used as a screening test for retinal abnormalities detection.
Keywords: local ternary pattern; LTP; local binary pattern; LBP; random forest; computer aided diagnosis; fundus images; K-nearest neighbourhood; KNN; diabetic retinopathy.
DOI: 10.1504/IJBET.2020.112421
International Journal of Biomedical Engineering and Technology, 2020 Vol.34 No.4, pp.334 - 353
Received: 06 Jul 2017
Accepted: 01 Dec 2017
Published online: 15 Jan 2021 *