Chest X-Ray Dataset With Annotations For Abnormality Detection

Vindr_Physionet_384 is a collection of chest X-ray images and associated annotations for machine learning in medical imaging. The dataset's 384 images are likely sourced from the VinDr-CXR and PhysioNet platforms, providing a labeled resource for computer vision tasks. It was created for research purposes, though the specific authors and update date are not provided.

Use Cases

• Train a convolutional neural network to classify chest X-ray images as normal or abnormal using image pixel data.
• Develop object detection models to localize specific findings like nodules or consolidations using bounding box annotations.
• Perform multi-label classification to predict the presence of multiple pathologies from image features and associated tabular labels.

Strengths

• Contains 384 medical images for model training and validation.
• Integrates data from two established medical imaging sources, VinDr-CXR and PhysioNet.

Limitations

• The dataset size of 384 images is relatively small for training deep learning models without augmentation.
• Specific annotation details, label distributions, and potential class imbalances are unknown.

Provenance

Source

Likely combines data from the VinDr-CXR and PhysioNet repositories.

Collection Method

Curated collection of chest X-ray images and annotations from open medical datasets.