Breast and Lymph Node Ultrasound Images with Expert Annotations for Detection Benchmarking

100 anonymized breast and axillary lymph node ultrasound images from Taichung Veterans General Hospital, released for public illustration and methodological reference. The dataset includes annotation files in JSON format for approximately 3,800 images, containing expert-defined lesion annotations and metadata. It was authored by Meng-Hsiun Tsai and last updated in April 2026.

Use Cases

• Benchmarking lesion detection algorithms based on the expert-defined annotations.
• Training image segmentation models based on the provided lesion boundary annotations.
• Developing workflows for medical image analysis based on the structured JSON annotation files.
• Studying methods for privacy-preserving data sharing in medical imaging based on the anonymization process.

Strengths

• Includes 100 curated, anonymized ultrasound images for public use.
• Provides expert-defined lesion annotations for approximately 3,800 images in a structured JSON format.
• Released under a CC-BY-4.0 license, facilitating open research and reuse.

Limitations

• Row count for the full curated image set is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The 30.7 MB size suggests a relatively small sample of images for model training.

Provenance

Source

Taichung Veterans General Hospital

Collection Method

Collected and anonymized in accordance with institutional privacy requirements and an IRB-approved Data Use Agreement.

Time Range

null

Freshness

Last updated 2026-04-24 15:45:49; freshness should be verified.

Geography

null