Table 3 Unsigned errors (in pixels) between the gold standard and each observer’s grid-manual and semi-automatic segmentation as well as the intra-observer errors.

From: Livelayer: a semi-automatic software program for segmentation of layers and diabetic macular edema in optical coherence tomography images

 

ILM

IPL–INL

OPL–ONL

Gold-observer1 (grid)

0.67

1.29

2.38

Gold-observer2 (grid)

0.74

1.02

1.82

Gold-observer3 (grid)

0.8

0.88

1.03

Gold-observer1 (semi)

2.75

1.86

1.85

Gold-observer2 (semi)

2.7

1.9

2.24

Gold-observer3 (semi)

2.65

1.94

2.11

Intra-observer (repeatability)

0.37

0.74

1.02