A review on image-based approaches for breast cancer detection, segmentation, and classification

The breast cancer as the most life-threatening disease among the woman has emerged in the worldwide. It is supposed that the early testing and treatment for breast cancer detection would be avoided the surgeries and increase the survival rate. A variety of research studies have motivated to improve the diagnostic methods for early diagnosis of breast cancer. This study investigates the automatic and semi-automatic image-based approaches for breast cancer diagnosis. The scope of this research has limited to the images based diagnosis application journal that are published between 2016 and 2020 years. The principles and associated risk factors for diagnosis the breast cancer and existing imaging techniques are presented. The steps of diagnosis including preprocessing, segmentation, extracting tumor features, and tumor classification are investigated. The publicly available datasets for breast imaging are briefly introduced as well. The application issues, challenges of breast imaging technologies and future directions are discussed. Based on the detailed study, most proposed methods use one type of imaging modalities, however, the doctor need to investigate the multiple imaging techniques to accurate diagnosis and effective treatment. Moreover, handling the multiple imaging require the processing of big data using a cluster computing framework.