MoTDiff: High-resolution Motion Trajectory estimation from a single blurred image using Diffusion models

Researchers have made a significant breakthrough in motion estimation with their new method called MoTDiff, which allows for high-resolution motion trajectory estimation from a single blurred image. This advancement is crucial for improving accuracy in various applications within computational imaging and computer vision, addressing the limitations of existing methods that often produce low-quality results. By leveraging diffusion models, this innovative approach promises to enhance the quality of motion information extraction, making it a noteworthy development in the field.