Skip to main content

Table 1 Classification and misclassification rates produced by the Algorithm and SR Classifications

From: A geometric method for computing ocular kinematics and classifying gaze events using monocular remote eye tracking in a robotic environment

Variables

Algorithm Classification

SR Classification

FQnS

90.7 (70.6, 99.6)

5.6 (0, 44.1)

SQnS

92.9 (39.6, 100)

76.2 (22.4, 90.1)

SPQnS

87.7 (56.1,100)

NA

misFQnS

6.4 (0.3,28.6)

1.17 (0, 3.8)

misSQnS

4.7 (0, 30.1)

0.07 (0, 1.72)

misSPQnS

3.0 (0.5, 12.0)

NA

  1. The gaze events computed using Algorithm Classification correctly classified around 90 % of fixations, saccades and pursuits. The incorrect classification rate of Algorithm Classification was around 5 %. These values are comparable with previously reported values for stimuli presented in vertical planes [42]. In contrast, SR Classification performed substantially worse at correctly identifying gaze events. Data are presented as mean (min, max)