Lumbar Coordinate Pretraining Dataset
This medical dataset contains 2,477 images of the lumbar region with precise annotations of 5 key points corresponding to the lumbar vertebrae, intended for pre-training for the degenerative classification of the spine.
2,477 medical images annotated with 5 key points, converted DICOM image format or PNG/JPEG possible
Attribution 4.0 International (CC BY 4.0)
Description
The “Lumbar Coordinate Pretraining Dataset” contains annotated medical images of the lower back (lumbar region). Each image contains five key points that indicate the position of the lumbar vertebrae, as well as enhanced sagittal coordinates for the left side of each lumbar disc, providing useful orientation information.
What is this dataset for?
- Pre-training medical imaging models for lumbar degenerative classification
- Development of algorithms for the automatic analysis of vertebral structures
- Improving model performance for identifying key anatomical points
Can it be enriched or improved?
The dataset can be supplemented with additional annotations, such as clinical classifications or images from other centers to increase diversity. 3D or temporal data could also enrich its use.
🔎 In summary
🧠 Recommended for
- Medical imaging researchers
- Health data scientists
- Medical computer vision projects
🔧 Compatible tools
- PyTorch
- TensorFlow
- MONAI
- ITK-SNAP
- LabelMe
💡 Tip
Use data augmentation techniques specific to medical imaging to improve the robustness of the model.
Frequently Asked Questions
Does this dataset contain 3D images or only 2D images?
This dataset mainly contains 2D images annotated with key points.
Can this dataset be used to train a degenerative classification model?
Yes, it is specially designed for pre-training in lumbar degenerative classification.
Does the license allow commercial use?
Yes, the CC BY 4.0 license allows commercial use under the condition of citation.
