A Framework for Promoting Passive Breast Cancer Monitoring: Deep Learning as an Interpretation Tool for Breast Thermograms

Document Type : Original Paper


1 Electrical Engineering Department, Iranian Research Organization for Science and Technology, Tehran, Iran

2 Department of Biomedical Engineering, Iranian Research Organization for Science & Technology (IROST), Tehran, Iran

3 Biomaterials and Tissue Engineering Department, Breast Cancer Research Center, Motamed Cancer Institute, ACECR, Tehran, Iran

4 Breast Diseases Department, Motamed Cancer Institute, South Gandhi St., Tehran, IR Iran.

5 Iranian Centre for Breast Cancer (ICBC), ACECR, Tehran, Iran

6 Genetics Department, Breast Cancer Research Center, Motamed Cancer Institute, ACECR, South Gandhi, Vanak Square, Tehran 1517964311, Iran.

7 Department of Computer Engineering, Faculty of Engineering, Islamic Azad University E-Campus, Tehran, Iran

8 Department of Computer Engineering, Faculty of Engineering, Alzahra University, Tehran, Iran


Introduction: Several types of cancer can be detected early through the advancement of thermography, a method of imaging tissues using thermal profiles. Thanks to its non-invasive, and radiation-free nature in parallel with its affability, thermography has gained increasing attention in recent years. There is a growing need for thermographic images of breast cancer lesions in different nationalities and ages to develop this technique, however.

Material and Methods: In this study, thermographic images of breast cancer from Iranian samples were prepared and confirmed due to the limited number of breast thermogram databases. To determine the reliability of the prepared database, in this research artificial intelligence is used to analyze the images from this database and a well-known DMR (Database for Mastology Research) database used in other studies.

Results: These databases are evaluated by using several deep learning architectures and transfer learning in terms of accuracy, sensitivity, speed, compliance with the training and validation process, as well as other factors. The results confirm the acceptable quality of the database obtained from this study in comparison with the DMR reference database, in such a way that the minimum accuracy, sensitivity, specificity, precision, and F-score of the above cases were equal to 80%, 86%, 86%, 88%, and 87%, respectively, according to best-fitted structures for both types of databases.

Conclusion: Using thermography as a method of early breast screening is demonstrated to be effective. However, the lower statistics on the study database (i.e., between 5 and 7 percent), show that more diverse breast thermograms need to be captured alongside improvements to imaging hardware and adherence to thermography recording protocols in order to increase database reliability and efficiency.


Main Subjects

Articles in Press, Accepted Manuscript
Available Online from 02 July 2023
  • Receive Date: 10 April 2023
  • Revise Date: 19 June 2023
  • Accept Date: 02 July 2023