Related Experiment Video
Updated: Dec 29, 2025
13:44
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
43.5K
Multi-View Mammographic Density Classification by Dilated and Attention-Guided Residual Learning.
IEEE/ACM Transactions on Computational Biology and Bioinformatics
|February 4, 2020
Summary
This study introduces an advanced radiomics method using deep learning for mammographic density classification, improving accuracy over existing techniques. The model shows promise for early breast cancer detection in computer-aided diagnosis systems.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Mammographic density is a key indicator for early breast cancer risk.
- Current classification methods often lack accuracy or require manual input.
Purpose of the Study:
- To develop an automated, accurate mammographic density classification system.
- To enhance early breast cancer risk assessment using deep learning.
Main Methods:
- A radiomics approach utilizing dilated and attention-guided residual learning.
- Implementation of a multi-stream network architecture for analyzing multi-view mammograms.
Main Results:
- Achieved 88.7% accuracy on a clinical dataset and 70.0% on a public dataset.
- Outperformed naive residual networks and recent deep learning approaches.
- Multi-view inputs improved classification accuracy by at least 2.0%.
Conclusions:
- The proposed deep learning model offers a significant advancement in mammographic density classification.
- The method demonstrates potential for integration into computer-aided diagnosis systems for improved breast cancer risk assessment.
