Meshcapade uses computer vision and machine learning to create lifelike digital humans from images, videos, or text.