A Comprehensive Study on Medical Image Denoising using Convolutional Neural Networks

Medical images are considered as one of the most important medical data for diagnosis and treatment purposes. For remote consultation and diagnosis, the medical images are frequently transferred using different communication methods. There is a chance of noise degradation of these medical images during transmission. Denoising such medical images is crucial for enhancing the quality and diagnostic precision. The remarkable ability of Convolutional Neural Networks (CNNs) to effectively denoise medical images has attracted considerable interest. This paper discusses various CNN architectures, training strategies, and data augmentation techniques for denoising tasks. In addition, the difficulties and limitations of CNN-based denoising techniques, such as computational complexity and data scarcity, are discussed. In addition, incorporating Generative Adversarial Networks (GANs) and transfer learning for improved denoising performance are discussed. This survey aims to serve as a valuable resource for researchers, clinicians, and developers by fostering a better comprehension of the current state-of-the-art techniques and future directions in medical image denoising using CNNs.