RUMPL: Ray-Based Transformers for Universal Multi-View 2D to 3D Human Pose Lifting

This article introduces RUMPL, a novel approach using ray-based transformers for 2D to 3D human pose estimation from multiple views. The use of transformers suggests an attempt to capture complex relationships within and between views. The 'universal' aspect implies a focus on broad applicability. The ArXiv source indicates this is a research paper, likely detailing the methodology, experiments, and results.