Making Artificial Intelligence systems robustly perceive humans remains one of the most intricate challenges in computer vision. Among the most complex problems is reconstructing 3D models of human hands, a task with wide-ranging applications in robotics, animation, human-computer interaction, and augmented and virtual reality. The difficulty lies in the nature of hands themselves, often obscured while holding objects or contorted into challenging orientations during tasks like grasping.