These kinds of projects usually involve three main steps. These three steps are Preprocessing, Feature Extraction and Feature Matching.
1. Preprocessing: Let's say if I am trying to classify ECG signal, I first need to remove noise from the signals and make them in compatible with my processing method.
2. Feature Extraction: Every data has its own distint features that help in classification. Find out the features that differentiate mammograms.
3. Feature Matching: One you are done with feature Extraction, you will match features with methods like SVM, ANN etc.
Now starting searching research papers concerned with your projects, find alogorithm from those papers and code it up.