Survey of Generative Approaches in Motion Generation

This document is a survey that categorizes and analyzes recent advancements in motion generation methods, focusing on generative approaches. Motion generation involves synthesizing realistic motion sequences from various inputs and is crucial in fields such as computer vision, graphics, and robotics. The survey reviews methods published in top-tier venues since 2023, providing a structured overview of generative strategies including GANs, autoencoders, and diffusion-based techniques. It details architectural principles, conditioning mechanisms, and generation settings, while compiling evaluation metrics and datasets used in the literature. The objective is to facilitate clearer comparisons among methods, identify open challenges, and serve as a foundational reference for researchers and practitioners. The paper also discusses the importance of understanding motion data representation and collection methods, emphasizing their impact on model design and performance. Overall, this survey aims to support both newcomers and experienced researchers in navigating the evolving landscape of motion generation research.