MURA Dataset: Towards Radiologist-Level Abnormality Detection in Musculoskeletal Radiographs.

We introduce MURA, a large dataset of musculoskeletal radiographs containing 40,895 images from 14,982 studies, where each study is manually labeled by radiologists as either normal or abnormal. On this dataset, we train a 169-layer densely connected convolutional network to detect and localize abnormalities. Six board-certified radiologists provide additional labels for a test set of 209 studies, on which we compare model and radiologist performance. We find that our model achieves performance comparable to that of radiologists. On finger, hand, and wrist studies, our modelu0027s F1 scores are slightly higher than those of radiologists, but not statistically significantly so; on elbow, forearm, humerus, and shoulder studies our modelu0027s F1 scores are slightly lower than those of radiologists, but also not statistically significantly so, indicating that the dataset presents a good challenge problem for future research. The dataset is freely available at this https URL