Table 8 \(M_{MAnnot}\) AUC for each CXR equipment on CHVNGE for each cross validation fold. Bold indicates the highest AUC for each fold.
From: Assessing clinical applicability of COVID-19 detection in chest radiography with deep learning
FUJI CR | Samsung | Carestream | FUJI DX |
---|---|---|---|
0.7808 | 0.5618 | 0.8056 | 0.6768 |
0.8145 | 0.5710 | 0.8551 | 0.7053 |
0.7770 | 0.5676 | 0.7959 | 0.7061 |
0.7908 | 0.5934 | 0.8249 | 0.7291 |
0.7975 | 0.5890 | 0.8019 | 0.6783 |