Chest X-Ray Images with Multimodal Medical Data

A multimodal dataset from ACM MM 2025, linking chest X-ray images to other medical information. Image processing requires padding the shorter side with zeros and resizing to 336 × 336 pixels. The dataset is authored by BoKelvin and was last updated in August 2025.

Use Cases

• Fine-tune vision models on chest X-ray images for medical image classification tasks.
• Develop multimodal models using linked image and text data for medical report generation.
• Analyze patterns in medical imaging data tagged with categories like 'Chest X Ray' and 'Medical Imaging'.

Strengths

• Dataset is associated with a peer-reviewed ACM MM 2025 conference track.
• Provides specific image preprocessing instructions (resize to 336 × 336).
• Includes multimodal tags indicating both image and text data components.

Limitations

• No information on dataset size, row count, or column structure is provided.
• Image data must be sourced separately from the MIMIC-CXR-JPG repository.
• Potential for incorrect data, as the author requests notification of errors.

Provenance

Source

MIMIC-CXR-JPG repository on PhysioNet for images.

Collection Method

Derived or linked from the MIMIC-CXR-JPG dataset; specific collection method for other data is unknown.

Freshness

Last updated on 2025-08-29.