CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning
About
We develop an algorithm that can detect pneumonia from chest X-rays at a level exceeding practicing radiologists. Our algorithm, CheXNet, is a 121-layer convolutional neural network trained on ChestX-ray14, currently the largest publicly available chest X-ray dataset, containing over 100,000 frontal-view X-ray images with 14 diseases. Four practicing academic radiologists annotate a test set, on which we compare the performance of CheXNet to that of radiologists. We find that CheXNet exceeds average radiologist performance on the F1 metric. We extend CheXNet to detect all 14 diseases in ChestX-ray14 and achieve state of the art results on all 14 diseases.
Pranav Rajpurkar, Jeremy Irvin, Kaylie Zhu, Brandon Yang, Hershel Mehta, Tony Duan, Daisy Ding, Aarti Bagul, Curtis Langlotz, Katie Shpanskaya, Matthew P. Lungren, Andrew Y. Ng• 2017
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Multi-Label Classification | ChestX-Ray14 (test) | AUROC (%)84.4 | 88 | |
| Thoracic Disease Classification | NIH ChestX-ray14 (test) | -- | 44 | |
| Thoracic Pathology Classification | ChestX-ray14 2017 (test) | -- | 29 | |
| Thoracic Disease Classification | MIMIC-CXR (test) | Atelectasis AUC56.7 | 28 | |
| Pathology Detection | NIH ChestX-ray 8 (test) | mAP (IoU@10-70)5.8 | 14 | |
| Pneumonia Detection | ChestX-ray14 2017 (test) | F1 Score43.5 | 6 | |
| Chest X-ray classification | OPI 4 (test) | Atelectasis AUC86.97 | 5 | |
| Chest X-ray classification | PDC 1 (test) | Atelectasis AUC84.99 | 5 | |
| Multi-label Chest X-ray Classification | OPI 4 (test) | Cardiomegaly AUC84 | 5 | |
| Multi-label Chest X-ray Classification | PDC 1 (test) | Cardiomegaly AUC0.8 | 5 |
Showing 10 of 11 rows