Harnessing Deep Learning (DL) for Image Inpainting in Healthcare System-Methods and Challenges
G. Sumathi, D. Uma · 2024
Over the past several years, technological breakthroughs have driven computer vision to become increasingly influential and valuable in various sectors. Computer vision has revolutionized the medical field by achieving many milestones that led to the automation of disease analysis. Many vision techniques have been used in medical imaging. One such prominent technique is image inpainting. Inpainting is a procedure of recovering the degraded regions of an image. Inpainting is also used to remove objects from an image. Recently, inpainting has become an eminent preprocessing technique in medical imaging while dealing with a subject such as a brain, chest, and liver. In medical imaging, abnormalities such as lesions, tumors, and abnormal tissues act as a barrier in disclosing general information about the subject from a set of victims. As a solution, inpainting is used to remove these lesions and tumors. The presence of metallic implants in X-ray images may obstruct the diagnosis process, which can also be solved by removing the obstructions using the inpainting process. In medical imaging, image inpainting is carried out by using various deep learning (DL) methodologies such as convolutional neural networks (CNNs) and generative adversarial networks (GANs). Thus, this chapter concentrates on the various methods of inpainting applied in various medical imaging techniques such magnetic resonance imaging (MRI), computed tomography (CT) scans, positron emission tomography (PET) scans, ultrasound imaging, X-ray imaging, etc., Furthermore, the problems and challenges in inpainting are also mentioned for future research scope.