Archives

Systematic Review on Machine Learning and Deep Learning Approaches for Mammography Image Classification


Saruchi and Dr. Surender Singh
Abstract

Early identification of bosom malignant growth is significant and exceptionally important in clinical practice. X-beam mammography is extensively utilized for prescreening the bosom and is likewise appealing because of its non-invasive nature. Early finding is crucially critical to build up a powerful treatment technique that will possibly lessen the death pace of a disease. Different tests can be considered to analyze and diagnose the breast malignant growth including examining of breast, breast ultrasound, magnetic resonance imaging, mammogram, and biopsy. In this astonishing methodology, impact on the classification of an image explicitly mammographic image classification is presented. In various methodologies, the challenge is to discover the region of interest (ROI) utilizing the approach of image processing. Subsequently, ROI next challenging test is the extraction of features and afterward it involves learning of various classes of images and prior detection methods of disease discovery. The result section involves the analysis of various methodological results and provides advance deep (or profound) learning approach over classification of mammography. In this audit, the classification methods of ROI, which work on a segmented or predefined ROIs with an attention over classification of mass are reviewed. An aggregate of 60 top notch conference and journal papers are chosen from various research database that meet a few incorporation rules. A relative examination is given dependent on ROI extraction strategies, machine learning methods and datasets are utilized along with prediction based on accuracy, and practice of frequency-based statistics.

Volume 12 | Issue 7

Pages: 337-350

DOI: 10.5373/JARDCS/V12I7/20202015