Multi Person 3D Pose Estimation

Multi-person 3D pose estimation aims to accurately determine the three-dimensional positions of multiple individuals' body joints from various input sources like RGB images, depth cameras, and LiDAR. Current research emphasizes robust performance in challenging scenarios, such as occlusion and varying viewpoints, using techniques like graph neural networks, transformer architectures, and adaptive camera selection for energy efficiency. These advancements are crucial for applications ranging from human-computer interaction and virtual reality to activity recognition and sports analysis, driving progress in both computer vision and related fields.

Papers