A Review on Recent Developments for Detection of Diabetic Retinopathy

Table 8

Different methods for detection of blood vessels.

Algorithm	Image processing techniques	Database	Color space	Sensitivity	Specificity	Accuracy
Bhatia et al. [28]	Gaussian filter, Canny edge detection, morphological operations, and Otsu thresholding	DRIVE; STARE	Green channel	70.31%	97.35%	95.23%
Fraz et al. [78]	Ensemble system of bagged and boosted decision trees, Gabor filter, and morphological transformation	STARE; DRIVE; CHASE_DB1	Green channel	0.75; 0.74; 0.72	0.97; 0.98; 0.97	0.95; 0.94; 0.94
Nguyen et al. [79]	Vessel segmentation based on the line detectors at varying scale	STARE; DRIVE	Green channel	—	—	0.93; Acc; 0.94; Acc
Fraz et al. [80]	Multiscale line detection method, Gabor filter	DRIVE; STARE; Messidor	Gray scale	0.73; 0.73; 0.77	0.97; 0.97; 0.98	0.94; 0.95; 0.96
Yin et al. [81]	Spectral clustering technique based on morphological features, Hessian matrix	DRIVE; REVIEW	Green channel	—	—	Above; 94%
Vega et al. [82]	Lattice Neural Networks with Dendritic Processing	STARE	Green channel	—	—	99.8%
Marín et al. [83]	Neural Network	DRIVE; STARE	Green channel	—	—	0.95; 0.97
Hou [84]	Multidirectional morphological top-hat transform, rotating structuring element	DRIVE; STARE	Green channel	0.73; 0.73	0.96; 0.96	0.94; 0.93
Shami et al. [85]	Morphological operations	Nikookari	Green channel	85.82%	99.98%	—