Adverse Weather Removal

Adverse weather removal aims to computationally restore clear images from those degraded by various weather conditions like rain, snow, and haze. Current research focuses on developing "all-in-one" models capable of handling multiple weather types simultaneously, often employing transformer-based architectures, mixture-of-experts models, or diffusion models to achieve this. These advancements are driven by the need for robust image processing in applications like autonomous driving and remote sensing, where diverse and unpredictable weather conditions are common. The development of efficient and accurate all-in-one models represents a significant step forward in computer vision and image processing.