Camera Keying

A technique that separates a person from the video background and generates an alpha matte without requiring a green screen. It is typically implemented frame by frame using deep learning-based human semantic segmentation models and is relatively sensitive to lighting changes and motion blur.