Motion Pattern
Motion pattern analysis focuses on understanding and modeling the movement of objects, particularly humans and robots, across various contexts. Current research emphasizes developing robust methods for predicting, generating, and classifying motion patterns using diverse techniques, including diffusion models, graph neural networks, and recurrent neural networks, often incorporating large language models for enhanced versatility and interpretability. This field is crucial for advancing robotics, autonomous systems, and human-computer interaction, as well as providing insights into human behavior and urban planning through improved trajectory prediction and anomaly detection.
Papers
MotionWavelet: Human Motion Prediction via Wavelet Manifold Learning
Yuming Feng, Zhiyang Dou, Ling-Hao Chen, Yuan Liu, Tianyu Li, Jingbo Wang, Zeyu Cao, Wenping Wang, Taku Komura, Lingjie Liu
Multi-Resolution Generative Modeling of Human Motion from Limited Data
David Eduardo Moreno-VillamarĂn, Anna Hilsmann, Peter Eisert