Machine Interaction-Based Computational Tools in Cancer Imaging
Praveen Kumar Gupta, Anushree Vinayak Lokur, Shweta Sudam Kallapur, Ryna Shireen Sheriff, Manjunatha Reddy, V. Chayapathy, Sindhu Rajendran, E. Keshamma
Abstract
Artificial intelligence (AI) is a field in which machines are engineered to become capable of being able to analyze real-time data and developing logical conclusions independently. AI gathers its own information and acts on its own to achieve the best outcome. Most, if not all, computational tools use AI as a means to carry out the respective processes. In particular, computational tools backed by algorithms driven by AI have gained major traction to become the focal point of healthcare in the 21st century. Primitive tools used to check the body for any abnormalities have been replaced by smart tools with higher computational power, greater accuracy, and higher rates of success. Cancer is a major cause of mortality among humans even in the modern age, with the lack of precise medicines to treat it in the later stages compounding the number of fatalities. The first step toward treating cancer in a patient is to visualize the location and aggressiveness of the tumor, which can be achieved using imaging, scientifically referred to as all “oncological imaging” or cancer imaging in colloquial terms. This chapter provides a comprehensive review of the various computational tools available to detect different types of cancers, as well as the AI/machine learning algorithms based on which they work. In particular, the imaging techniques delineated in this review include the temporal subtraction method, segmentation techniques, feature extraction methods, and network training. Lung, skin, bone, brain, and breast cancers, as well as the cancer of the immune system (lymphoma), have been focused upon and the machine learning algorithms used to detect them have been elaborated. The intricacies of these algorithms, such as artificial neural networks (ANN) and deep learning, have also been discussed and broken down into segments. This chapter provides an extensive review of the computational tools used for modern cancer imaging, as well as information about the intersection of AI and machine learning that occurs in order to ensure the proper working of the tools, to detect cancer as quickly and accurately as possible.