Skip to main content

Table 1 Description of the designed input feature maps

From: Improving the performance of deep learning models in predicting and classifying gamma passing rates with discriminative features and a class balancing technique: a retrospective cohort study

ID

Input feature map

Description

F1

Fluence

The planned fluence to be delivered to each pixel in the matrix, including that delivered through the leaf gap opening, transmitted through the MLC leaves, and transmitted through the jaw collimators

F2

Gap

The planned average width of the leaf gaps passing each pixel in the matrix

F3

SpeedA

The planned leaf speed for bank A leaves passing each pixel in the matrix

F4

SpeedB

The planned leaf speed for bank B leaves passing each pixel in the matrix

F5

DiffFluence

The difference between the actual and planned fluence delivered to each pixel in the matrix

F6

DiffGap

The difference between the actual and planned average width of the leaf gaps passing each pixel in the matrix

F7

DiffSpeedA

The difference between the actual and planned leaf speed for bank A leaves passing each pixel in the matrix

F8

DiffSpeedB

The difference between the actual and planned leaf speed for bank B leaves passing each pixel in the matrix