Table 1 Overview of the inputs to the classification algorithm used per method in the evaluation of this study.

From: Glaucoma detection using non-perfused areas in OCTA

Approach

Input to classification

CNN

Image

VD

Percentage

GPDP

Percentage

IA

Maximum

Mean

Standard deviation

Kurtosis

PD

Maximum

Mean

Standard deviation

Kurtosis

PDA

Maximum

Mean

Standard deviation

Kurtosis

  1. CNN convolutional neural network, VD vessel density, GPDP geometric perfusion deficite percentage, IA intercapillary areas, PD perfusion distance, PDA perfusion deficite areas.
Back to article page