Computational Intelligence Methods for Super-Resolution in Image Processing Applications | 2021
A Comprehensive Review of CAD Systems in Ultrasound and Elastography for Breast Cancer Diagnosis
Abstract
Since the causes of breast cancer remain unknown, early diagnosis can increase the survival rate and reduce the mortality rate. Screening is a powerful way to detect breast cancer (BC). Screening methods, including mammography, sonography, and elastography, are commonly utilized for early BC diagnosis. However, the medical image’s manual analysis is a highly challenging and time-consuming process and often leads to a disagreement between radiologists. Recently, computer-aided diagnosis (CAD) systems proved the potential for detecting and classifying BC with high accuracy. This text aims to assess the performance of ultrasound and elastography for the BC early detection. Many problems involved in BC detections arise, and different approaches, along with their strengths and drawbacks, are investigated. This survey also explores the selection of input features, classification techniques adopted, and performance indices used in each research work. From the examined literature, it is noticed that the artificial neural network (ANN) can support radiologists to make an accurate decision. Numerous imaging techniques help the physician’s decision-making at several theragnostic stages (including diagnosis evaluation, treatment choice, interventional assistance, and follow-up). This active research subject entails many efforts to exceed the current pixel resolution to the molecular level in several imaging modalities. The present usage and future usage of high-resolution (HR) images depend on intrinsically analyzing a massive number of images that are not easy to manage and process by either radiologists or surgeons.