[Retracted] Research on Chest Disease Recognition Based on Deep Hierarchical Learning Algorithm

Table 1

Summary of 14 classes in CheXPert dataset (https://stanfordmlgroup.github.io/competitions/chexpert/).

Pathology	Positive	Negative	Uncertain	Empty	Pathology	Positive	Negative
No finding	22381	0	0	201033	No finding	38	196
Enlarged cardiom.	10798	21638	12403	178575	Enlarged cardiom.	109	125
Cardiomegaly	27000	11116	8087	177211	Cardiomegaly	68	166
Lung opacity	105581	6599	5598	105636	Lung opacity	126	108
Lung lesion	9186	1270	1488	211470	Lung lesion	1	233
Edema	52246	20726	12984	137458	Edema	45	189
Consolidation	14783	28097	27742	152792	Consolidation	33	201
Pneumonia	6039	2799	18770	195806	Pneumonia	8	226
Atelectasis	33376	1328	33739	154971	Atelectasis	80	154
Pneumothorax	19448	56341	3145	144480	Pneumothorax	8	226
Pleural effusion	86187	35396	11628	90203	Pleural effusion	67	167
Pleural other	3523	316	2653	216922	Pleural other	1	233
Fracture	9040	2512	642	211220	Fracture	0	234
Support devices	116001	6137	1079	100197	Support devices	107	127