Skip to main content
. 2023 Aug 24;36(6):2402–2410. doi: 10.1007/s10278-023-00894-x

Table 2.

Label distributions of images in January to February, including TRAIN_PRI, VAL_EVAL, and VAL_PTEST

Dataset Normal no. (%) Abnormal no. (%) Arthroplasty no. (%) Total no.
TRAIN_PRI 198 (23%)  589 (69%)  67 (8%)  854
VAL_EVAL 41 (20%)  152 (73%)  15 (7%)  208
VAL_PTEST 44 (21%)  142 (68%)  22 (11%)  208
TRAIN_SECa 1867 (28%)  4282 (64%)  507 (8%)  6656
TEST 151 (23%)  457 (69%)  51 (8%)  659

aLabels in TRAIN_SEC were pseudo-labeled by the NLP model