Skip to main content

Table 6 Descriptions of features for MA detection

From: Microaneurysms detection in color fundus images using machine learning based on directional local contrast

Feature types

Symbols

Descriptions

Color

f1~2

Mean and standard deviation value of candidate patch in RGB color

f3~4

Mean and standard deviation value of candidate patch in HSV color

f5~6

Mean and standard deviation value of candidate patch in CIElab color

Grayscale

f7~8

Mean and standard deviation value of candidate patch in \(I_\text {g}\)

f9~10

Mean and standard deviation value of candidate patch in \(I_\text {CLAHE}\)

DLC

f11~22

Directional local contrast (DLC) of the center point of each candidate region in \(I_\text {CLAHE}\)

Shape

f23~28

Area, Perimeter, Circularity, Eccentricity, Aspect ratio, and Solidity of each candidate region

Texture

f29~32

Entropy, Energy, Homogeneity and Skewness of candidate patch in \(I_\text {CLAHE}\)

Gaussian filter-based

f33~40

Mean and standard deviation value of candidate patch in corresponding Gaussian filtered result of \(I_\text {CLAHE}\) when \(\sigma\) is \(\left[ 1,2,4,8\right]\)

Gradient

f41~42

Mean gradient of candidate patch in \(I_\text {CLAHE}\) (mean value of dx and dy)

f43~44

Mean gradient on the boundary of each candidate region in \(I_\text {CLAHE}\) (mean value of dx and dy on the boundary)