Skip to main content

Table 2 The error categorization for images with poor results

From: Automating Rey Complex Figure Test scoring using a deep learning-based approach: a potential large-scale screening tool for cognitive decline

 

5 ≤ absolute difference

10 ≤ absolute difference

5 ≤ absolute difference < 10

Total

188 (100%)

13 (100%)

175 (100%)

Image quality

1 (0.5%)

0 (0%)

1 (0.6%)

Scoring

61 (32.4%)

1 (7.7%)

60 (34.3%)

Digitalization

18 (9.6%)

12 (92.3%)

6 (3.4%)

Model bias

108 (57.5%)

0 (0%)

108 (62.7%)