3D Pose

3D pose estimation aims to reconstruct the three-dimensional position and orientation of objects, primarily humans, from various input modalities like images or videos. Current research heavily focuses on improving accuracy and robustness, particularly in challenging scenarios involving occlusions, limited viewpoints, and complex interactions, employing techniques like transformer networks, diffusion models, and Gaussian splatting for representation and refinement. These advancements are crucial for applications ranging from augmented and virtual reality to robotics and healthcare, enabling more natural and accurate human-computer interaction and analysis of human movement.