Recognition and Prediction of Breast Cancer using Supervised Diagnosis

Abstract

Breast cancer is the most common and a major death causing disease diagnosed among women worldwide. Early detection of this disease can reduce the death rates. Image processing techniques using machine learning are widely used in medical domain to improve the early detection of cancerous tumors in breast. In this proposed approach, supervised learning techniques are used to extract cancer defining features and classify cancerous images from the normal mammogram images. The supervised system is initially trained by extracting 13 features each from a dataset of 30 images. The extracted features of the image under test are associated with the features extracted from the database images to detect and predict the cancer tumor in the image. Support Vector Machine (SVM) and K-Nearest Neighbours(KNN) is used for classification. Based on the analysis, the system is capable to give a classification accuracy of 95%(SVM) and 97% (KNN). A GUI based interface is also developed for the same. Further, a user-friendly chatbot is developed using Dialog Flow, which interacts with patients to predict cancer based on the symptoms identified by the patient. This chatbot can be used by the patient to detect whether the symptoms are porne to cancer.