Skip to main content
. 2024 Feb 26;11:242. doi: 10.1038/s41597-024-03000-0

Table 2.

Samples distribution of HITL iterations.

Sample types Training Test Sum
1st Iteration 2nd Iteration 3rd Iteration
Healthy 14 (100.0%) 7 (58.3%) 6 (50.0%) 3 30
Pathological 0 (0.0%) 5 (41.7%) 6 (50.0%) 2 13
Sum 14 12 12 5 43

Over the iterations, the percentage of pathological samples, which are difficult to learn compared to the healthy samples, was continuously increased.