Related Experiment Video
Updated: Jul 18, 2026
06:08
Quantitative Visualization and Detection of Skin Cancer Using Dynamic Thermal Imaging
Published on: May 5, 2011
DensiThAI: A Multi-View Deep Learning Framework for Breast Density Estimation Using Infrared Images
Siva Teja Kakileti1, Geetha Manjunath2
1Niramai Health Analytix Pvt. Ltd., Koramangala, Bangalore, 560095, Karnataka, India. sivateja@niramai.com.
Journal of Imaging Informatics in Medicine
|July 13, 2026
Summary
Infrared thermal imaging may detect breast tissue density patterns using artificial intelligence. This novel approach could complement mammography for breast cancer risk assessment.
Area of Science:
- Biomedical Imaging
- Artificial Intelligence in Medicine
- Breast Cancer Research
Background:
- Breast tissue density is a key breast cancer risk factor.
- Mammography is the standard for density assessment but has limitations.
- Fibroglandular and adipose tissues may have distinct thermal properties.
Purpose of the Study:
- To investigate if infrared thermal images can reflect breast tissue density information.
- To develop an AI framework for density classification using thermal imaging.
- To evaluate the potential of thermal imaging as a complementary breast imaging tool.
Main Methods:
- Proposed DensiThAI, a multi-view deep learning framework.
- Integrated data from five standardized thermal breast views.
- Evaluated on a multi-center dataset of 3500 women with mammography-derived density labels.
Main Results:
- DensiThAI achieved a mean Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.73.
- Demonstrated statistically significant separation between breast density classes (p < 0.05).
- Identified detectable density-associated thermal patterns.
Conclusions:
- Infrared thermal imaging shows potential for breast tissue characterization.
- AI analysis of thermal images may offer complementary information to mammography.
- Further validation in larger cohorts is warranted for clinical application.
