RTM: Anonymized Mammography Images for Breast Cancer Detection Research

A curated collection of anonymized mammographic breast imaging data developed for research in computer-aided breast cancer detection. The dataset contains preprocessed mammogram images in PNG format, totaling 4.3 MB, and includes representative benign, malignant, and normal breast cases. It was authored by Thirupathi Rao N and last updated on 2026-05-18.

Use Cases

• Training computer-aided breast cancer detection models based on preprocessed mammogram images.
• Developing lesion segmentation algorithms based on provided annotation and lesion mask information.
• Conducting multi-class mammogram classification (benign, malignant, normal) using transformer- and CNN-based architectures.
• Performing feature extraction for medical image analysis research based on the dataset's metadata and class distribution.

Strengths

• Dataset is preprocessed and curated for research tasks, including lesion localization and segmentation.
• Includes annotation and lesion mask information where applicable, supporting supervised learning.
• All personally identifiable patient information was removed, ensuring privacy protection and ethical compliance.
• Released under a permissive CC0-1.0 license for academic and non-commercial research.

Limitations

• Dataset size is 4.3 MB, which is small for image-based deep learning tasks.
• Row count and exact number of images are unknown, limiting suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

figshare

Collection Method

Curated collection of anonymized mammographic breast imaging data collected under institutional research guidelines.

Freshness

Last updated 2026-05-18 07:44:38; freshness should be verified.