AI/ML techniques in servicing LDCT reconstruction: a systematic literature review

The computed tomography (CT) scan is an extensively utilized technique for providing good visualization of human body parts. On the other hand, low-dose computed tomography (LDCT) has attracted researchers' interest in health imaging due to the potential dangers it poses in the form of body X-rays. To address these issues, researchers have devised a number of methods to enhance computed tomography image reconstruction. After conducting a thorough analysis of each record, taking into consideration the objectives and scope of the study, we chose 164 research publications to be included in this evaluation. The best algorithms for CT reconstruction found in this comprehensive literature are deep learning reconstruction (DL), convolutional neural networks (CNNs), generative adversarial networks (GANs), residual networks (RED-CNNs), Recurrent neural network (RNN), U-Net and others. As an outcome of this systematic survey, it has been observed that the application of deep learning may enhance CT image reconstruction efficacy, elevate image quality, and minimize radiation exposure. Diagnostic accuracy, and overall healthcare system efficiency can be accomplished through the utilization of CNN technology, which can make CT image reconstruction more precise and efficient.