Analysis of Visual Appearance of Retinal Nerve Fibers in High Resolution Fundus Images: A Study on Normal Subjects

Table 2

Definitions of the first-order features used for analysis.

Mean	𝜇=∑^𝐺−1𝑔=0𝑔𝐻𝑔	H(g) represents the probability density function, estimated from histogram 𝐻(𝑔)=𝑛𝑔/𝑁, where pixel value 𝑔=0,1,2,…,𝐺−1,𝐺 is a number of gray levels, 𝑁 is a number of pixels in analyzed image, and 𝑛𝑔 is a number of pixels with value 𝑔. 𝜇𝑛 represents statistical moment of nth order: 𝜇𝑛=∑^𝐺−1𝑔=0(𝑔−𝜇)^𝑛𝐻(𝑔)
Standard deviation	𝜎=∑^𝐺−1𝑔=0𝑔−𝜇^2𝐻𝑔
Shannon entropy	𝐸=−∑^𝐺−1𝑔=0𝐻𝑔log2𝐻𝑔
Skewness	𝛾1=𝜇3/𝜇2^(3/2)
Kurtosis	𝛾2=𝜇4/𝜇^22−3