Breast cancer screening application using thermographic images and machine learning analysis

• Main Authors: Evidente, Rency Christelle O., Sayoc, Vincent Rafael T., Uy, Jason Patrick T., Young, Leila Yasmin C.
• Format: text
• Language: English
• Published: Animo Repository 2022
• Subjects: Temperature measuring instruments; Breast—Cancer—Imaging; Biomedical; Electrical and Computer Engineering
• Online Access: https://animorepository.dlsu.edu.ph/etdb_ece/15; https://animorepository.dlsu.edu.ph/cgi/viewcontent.cgi?article=1026&context=etdb_ece
• Institution: De La Salle University

In the Philippines, breast cancer is one of the most prevalent cancers that affect women. The group developed an Android app that uses thermograms and machine learning analysis to determine whether the breasts are healthy or unhealthy in order to increase awareness and undertake early breast screenings for breast cancer symptoms. A breast thermogram dataset from Visual Lab was utilized for training and testing five different types of mobile-friendly convolutional neural networks in addition to creating this application. A network called EfficientNetB0 was selected since it consistently outperformed all other networks in tests. The network was then trained and tested using the TFLite Model maker. This network was incorporated into an Android application that could anticipate and annotate the data using the trained EfficientNetB0 and image analysis using a range of pixel values that are dominantly present in unhealthy conditions. The group created a survey to get input on the usability of the application's user interface from medical students. Despite the program's potential for successful operation, it cannot be put into use because it goes against the identical and independently distributed (i.i.d.) random variable assumption.