Comparative Review of Mammary Cancer using Machine Learning Techniques

The discovery of mammary cancer primarily depends on examining tissues slide through a microscope. This method is time consuming and subjects to human error. Various traditional machine learning techniques require specialists to manually select features such as texture, color or shapes. These aspects frequently don't work well for every image, especially when the slides have different hues or staining methods. This paper presents an automated system that employs on Convolutional Neural Networks (CNN) for the analysis of mammary cancer. This study is to develop a patch-based CNN framework capable of automatically analyze small parts of large breast tissue images and helps to identify whether they are benign or malignant. Now, instead of depending on manually selected features, system learns important patterns directly from the image patches. An accurate final diagnosis for the entire slide or patient is subsequently created by combining these patch-level findings. In order to demonstrate how much better and more dependable CNN is, the study also contrasts its performance with that of conventional techniques.