Low Light Video

Low-light video enhancement aims to improve the quality and visibility of videos captured in dimly lit conditions, addressing challenges like noise, low contrast, and temporal inconsistencies. Current research focuses on developing deep learning models, including convolutional neural networks, transformers, and diffusion models, often employing unpaired learning techniques and incorporating spatio-temporal alignment to leverage information across frames. These advancements are driven by the need for improved performance in computer vision applications like autonomous driving and surveillance, as well as the creation of high-quality, fully registered datasets to facilitate model training and evaluation.