The project description and results are discussed in the TAI_proj_report pdf file. The following text is the introduction of the report.
In this project machine-learning models will be used to try and detect breast cancer based on
X-ray images. On these images tumours are difficult to see due to the anatomy of the breast,
thus micro-calcifications (MCs or micros) are used to detect breast cancer. MCs are white
calcium deposits that are easy to see on X-ray images, unlike tumours, and can be indicative
of cancer as there is a correlation between MCs and tumours. Malignant MCs are MCs that
are in the neighbourhood of tumours, while benign MCs are not.
The project consists of first training a model that can accurately predict whether a micro is
malignant or benign. Afterwards, patients are classified whether they have cancer based on
the classifications of the micros.