Analyzing histopathological images by using machine learning techniques

Medical image data have become an important part of every patient's digital health record. With the advancement of microscope technology, pathologists can now handle histopathological tissue slides more quickly with digitized WSI. Manual evaluations of massive histological images are time taking and sometimes error-prone, particularly for pathologists with diverse degrees of skill. Patient can be harmed by a delayed or erroneous analysis. Our research work combines image processing techniques (grayscale, edge-detection) plus supervised machine learning algorithms such as RF, SVM, and KNN for analyzing histopathological images (HI) and finds the optimal algorithm to classify breast cancer. Breast cancer is the major malignant common cancer in women after lung cancer; it is the 2nd biggest cause of death from cancer of women. RF algorithm achieved 98.2 and 98.3% accuracy for Benign, and Malignant cancer compared with other algorithms to classify breast cancer on WSI dataset.