Soft tissue sarcoma diagnosis using machine and deep learning-survey

Abstract

The collection of unique and diverse tumors known as soft-tissue sarcomas (STS) is hindered by a number of factors, such as delayed or inaccurate diagnosis, and a lack of clinical knowledge, and a restricted range of treatment alternatives. The tissues that surround, link, and support other body organs and structures are the target of a rare type of cancer known as soft tissue sarcomas. Muscle, fat, blood vessels, deep skin tissues, tendons, and ligaments are among the tissues that can be impacted by soft tissue sarcomas. Soft tissue sarcomas can arise in nearly every body component, including the arms, legs, and abdomen. The way that patients are treated medically is severely harmed by these diagnostic mistakes. Numerous machine learning models have been proposed by researchers to categorize cancers, but none of them have sufficiently addressed the issue of misdiagnosis. Furthermore, the majority of comparable research that has suggested models for the assessment of these malignancies do not take the heterogeneity and volume of the data into account. This research presents the comparison between machine and deep learning methods for the improved categorization of soft tissue sarcomas. This research further proposes on the early detection of STS. In the next stage of classification, an optimal Convolution Neural Network (CNN) is employed.