RHD stands for Rendered Hand Pose, which is an image dataset for hand gesture estimation. The dataset contains 41,258 training sets and 2,728 test sets.

Each sample provides:

• RGB image (320 x 320 pixels)
• Depth map (320 x 320 pixels)
• Class segmentation mask (320 x 320 pixels): person, background, 3 classes for each finger, and 1 class for each palm
• 21 hand key points with uv coordinates in the image frame, xyz coordinates in the world frame, and a visibility indicator
• intrinsic camera matrix K