Motion Estimation
Motion estimation aims to determine the movement of objects or cameras within a scene, a crucial task across diverse fields like robotics, medical imaging, and video processing. Current research emphasizes improving accuracy and robustness, particularly in challenging scenarios involving complex motions, occlusions, and limited data, often employing deep learning models (e.g., diffusion models, transformers) and integrating multiple sensor modalities (e.g., LiDAR, IMU, event cameras). These advancements have significant implications for applications ranging from autonomous navigation and surgical robotics to medical image reconstruction and video compression, enabling more accurate and efficient systems.
Papers
Hard to Track Objects with Irregular Motions and Similar Appearances? Make It Easier by Buffering the Matching Space
Fan Yang, Shigeyuki Odashima, Shoichi Masui, Shan Jiang
The Second-place Solution for ECCV 2022 Multiple People Tracking in Group Dance Challenge
Fan Yang, Shigeyuki Odashima, Shoichi Masui, Shan Jiang
Motion estimation and filtered prediction for dynamic point cloud attribute compression
Haoran Hong, Eduardo Pavez, Antonio Ortega, Ryosuke Watanabe, Keisuke Nonaka
A Codec Information Assisted Framework for Efficient Compressed Video Super-Resolution
Hengsheng Zhang, Xueyi Zou, Jiaming Guo, Youliang Yan, Rong Xie, Li Song