Skip to main content

Table 6 Comparison of multiple label classification methods for thoracic diseases. The classification methods, measurements, and best results in the review are shown in each column, respectively

From: Computer-aided detection in chest radiography based on artificial intelligence: a survey

Methods Thoracic diseases Measurements Best results
RESNET [23] Atelectasis, cardiomegaly effusion, infiltration, mass, nodule, pneumonia, pneumothorax, consolidation, edema, emphysema, fibrosis, pleural thickening hernia AUC Respectively, 71.6%, 80.7%, 78.4%, 60.9%, 70.6%, 67.1%, 63.3%, 80.6%, 70.8%, 83.5%, 81.5%, 76.9%, 70.8%, 76.7%
LSTM + DENSENET [107] AUC Respectively, 77.2%, 90.4%, 85.9%, 69.5%, 79.2%, 71.7%, 71.3%, 84.1%, 78.8%, 88.2%, 82.9%, 76.7%, 76.5%, 91.4%
ChexNet [13] AUC Respectively, 82.1%, 90.5%, 88.3%, 72.0%, 86.2%, 77.7%, 76.3%, 89.3%, 79.4%, 89.3%, 92.6%, 80.4%, 81.4%, 93.9%
Cascade deep learning network based on DENSENET [110] AUC Respectively, 76.2%, 91.3%, 86.4%, 69.2%, 78.9%, 70.4%, 71.5%, 85.9%, 78.4%, 88.8%, 91.6%, 75.6%, 77.4%, 89.8%
Attention guided CNN [111] AUC Respectively, 85.3%, 93.9%, 90.3%, 75.4%, 90.2%, 82.8%, 77.4%, 92.1%, 84.2%, 92.4%, 93.2%, 86.4%, 83.7%, 92.1%