ABSTRACT
In early days, each and every field of medical image analysis and diagnosis took more time to conclude the medical report of patients. Most of the time, it resulted in failure, and it is a challenging task for the medical field. More than a decade’s gap between technology and human society creates drastic loss in human life. In those days, radiologists were not able to make successful attempts in radiology. In any field, technology is the path for evaluation. In the early computer age, the recital level of AI and human intelligence started and tried to bind one another. In the initial stage, AI started with little success and now it is evidently surpassing the level of human intelligence. AI grounds its foot stronger to the widespread evaluation of medical imaging. In this technological era, the medical field uses a variety of techniques to examine every inner organ for diagnosis and treatment of the human body. Every medical testing needs a different type of image analysis in the form of X-rays, DICOM images, MRI scans, etc. All these images have different qualities and nature depending on patients and illumination around them. Medical image diagnosis and identification is a roughly tougher task for radiologists due to the different nature and modality of images, and it takes more time for analysis in different platforms and techniques. To prevent and cure diseases, physicians need prolonged support from radiologists and they need the hands of computer-aided diagnosis (CAD) tools and techniques. Machine learning is a branch of AI. Deep learning is a subset of machine learning and both areas are twins of AI. They are inseparable and, at the same time, each technology explores its medical analysis in different ways. This chapter explores application of machine learning and deep learning in chest X-ray image analysis. There is no limit in the use of machine learning techniques. It branches its root to maximum areas in the development of medical field. This chapter presents various machine learning techniques for segmentation and classification with case study.
