Modulo Video Recovery via Selective Spatiotemporal Vision Transformer

The introduction of the Selective Spatiotemporal Vision Transformer (SSViT) marks a pivotal advancement in the field of video recovery, particularly for modulo cameras that have been around for over a decade. Traditional HDR methods have proven unsuitable for recovering folded video frames, which has hindered progress in this area. SSViT, as the first deep learning framework specifically designed for modulo video reconstruction, employs a novel token selection strategy to enhance efficiency and focus on critical regions of the video. This innovation not only addresses the limitations of existing methods but also achieves state-of-the-art performance in modulo video recovery, showcasing the potential of modern deep learning techniques in overcoming longstanding challenges in imaging technology.