Using a deep learning based human body frame detection model that is trained with massive dataset and machine learning processes, SenseTime can accurately estimate the position and posture of the human body in a single-frame RGB-D image and provide the corresponding 17 3D coordinates of the body joints. The technology is accurate, fast, and covers a wide range of common use cases. This includes: 3D human pose estimation: To provide 3D joint coordinates of human body frame based on RGB-D information input. Body segmentation with depth info: To segment the human body and the background and generate a depth image based on RGB-D information input.