Breast Cancer Diagnosis Using Image Processing and Machine Learning

Abstract

Breast cancer is one of the second leading causes of cancer death in women. Despite the fact that cancer is preventable and curable in primary stages, a huge number of patients are diagnosed with cancer very late. Conventional methods of detecting and diagnosing cancer mainly depend on skilled physicians, with the help of medical imaging, to detect certain symptoms that usually appear in the later stages of cancer. Therefore, we present here the computerized method for cancer detection in its early stage within a very short time. Here, we have used Machine learning to train a model using the predicted features of the nuclei of cells. A comparative study of two different algorithms KNN and SVM is conducted where the accuracy of each classifier is measured. After this, we analyze a digital image of a fine needle aspirate (FNA) of breast tissue using image processing to find out the features of nuclei of the cells. We then apply the feature values to our trained model to find whether the tumor developed is benign or malignant.