Artifact Removal

Artifact removal aims to cleanse images and signals from distortions that hinder analysis and interpretation, impacting diverse fields from medical imaging to brain-computer interfaces. Current research heavily utilizes deep learning, employing transformer networks, convolutional neural networks, and generative models like diffusion models and normalizing flows to address artifacts in various modalities, including EEG, CT scans, MRI, and even images from virtual try-on applications. These advancements improve diagnostic accuracy in medicine, enhance the reliability of neuroscientific studies, and improve the quality of image-based applications. The development of specialized datasets and evaluation metrics further strengthens the rigor and reproducibility of artifact removal research.