The system fills the detected face with archive images. Skin becomes the base layer; each eye, brow, nose, and mouth is matched and placed back onto the face. The text above each face is the system's compressed reading, not a real emotion.