The Visual Big Data Research Center studies virtual human synthesis technology, synthesizes realistic virtual human videos based on speech text and speech, and synthesizes automatic question-and-answer virtual humans in combination with large language models; on the basis of a series of face generation technologies, researches 3D human/face real-time drive technology; understands the emotions of the other party during the interaction process, and generates dialogue face responses with corresponding expressions.