Motion Information
Motion information research focuses on accurately estimating and utilizing movement data from various sources, including videos, sensor data, and medical images, for diverse applications. Current research emphasizes developing robust and efficient algorithms, often employing deep learning models like diffusion models and Siamese networks, to address challenges such as motion blur, occlusions, and limited training data. These advancements are significantly impacting fields like computer vision, robotics, and medical imaging, enabling improved 3D reconstruction, autonomous navigation, and medical image analysis. The development of more accurate and generalized motion models continues to be a key focus.
Papers
Deformable Convolutions and LSTM-based Flexible Event Frame Fusion Network for Motion Deblurring
Dan Yang, Mehmet Yamac
We never go out of Style: Motion Disentanglement by Subspace Decomposition of Latent Space
Rishubh Parihar, Raghav Magazine, Piyush Tiwari, R. Venkatesh Babu
Example-based Motion Synthesis via Generative Motion Matching
Weiyu Li, Xuelin Chen, Peizhuo Li, Olga Sorkine-Hornung, Baoquan Chen
QPGesture: Quantization-Based and Phase-Guided Motion Matching for Natural Speech-Driven Gesture Generation
Sicheng Yang, Zhiyong Wu, Minglei Li, Zhensong Zhang, Lei Hao, Weihong Bao, Haolin Zhuang
A Framework Based on Symbolic Regression Coupled with eXtended Physics-Informed Neural Networks for Gray-Box Learning of Equations of Motion from Data
Elham Kiyani, Khemraj Shukla, George Em Karniadakis, Mikko Karttunen
Motion Matters: Neural Motion Transfer for Better Camera Physiological Measurement
Akshay Paruchuri, Xin Liu, Yulu Pan, Shwetak Patel, Daniel McDuff, Soumyadip Sengupta
MSTFormer: Motion Inspired Spatial-temporal Transformer with Dynamic-aware Attention for long-term Vessel Trajectory Prediction
Huimin Qiang, Zhiyuan Guo, Shiyuan Xie, Xiaodong Peng