VinDr-RibCXR: A Benchmark Dataset for Automatic Segmentation and Labeling of Individual Ribs on Chest X-rays
This provides a benchmark for medical imaging researchers working on rib analysis in chest X-rays, but it is incremental as it focuses on dataset creation and baseline performance.
The authors tackled the problem of automatic segmentation and labeling of individual ribs on chest X-rays by introducing a new benchmark dataset, VinDr-RibCXR, and achieved a Dice score of 0.834 with their best model on a test set.
We introduce a new benchmark dataset, namely VinDr-RibCXR, for automatic segmentation and labeling of individual ribs from chest X-ray (CXR) scans. The VinDr-RibCXR contains 245 CXRs with corresponding ground truth annotations provided by human experts. A set of state-of-the-art segmentation models are trained on 196 images from the VinDr-RibCXR to segment and label 20 individual ribs. Our best performing model obtains a Dice score of 0.834 (95% CI, 0.810--0.853) on an independent test set of 49 images. Our study, therefore, serves as a proof of concept and baseline performance for future research.