Motion Reconstruction
Motion reconstruction aims to accurately capture and represent 3D human movement from various input sources, such as video, sparse sensor data (e.g., IMUs), or even audio. Current research heavily utilizes deep learning, employing architectures like diffusion models, transformers, and graph neural networks to address challenges like occlusion, noise, and sparse data. These advancements are improving the accuracy, robustness, and real-time capabilities of motion reconstruction, with significant implications for virtual reality, animation, healthcare, and human-computer interaction.
20papers
Papers
February 27, 2025
December 11, 2024
December 4, 2024
November 15, 2024
November 7, 2024
October 27, 2024
September 20, 2024
September 10, 2024
July 18, 2024
April 29, 2024
January 16, 2024
December 27, 2023
December 12, 2023