Revolutionising Engineering Education: Creating Photorealistic Virtual Human Lecturers Using Artificial Intelligence And Computer Generated Images

The COVID-19 pandemic has disrupted traditional classroom learning, making virtual and remote education increasingly important. In this context, the use of photorealistic virtual humans, or avatars, powered by Artificial Intelligence (AI) can offer an immersive and engaging environment for delivering traditional classroom-based lectures. This paper proposes a process that combines AI and Computer-Generated Images (CGI) to create photorealistic virtual human lecturers for educational purposes. The proposed process flow involves generating audio from text inputs, which is passed to a 3-Dimensional (3D) facial animation rig that matches lip, tongue, eye and facial movements to the audio using AI. This generates a base mesh for speech animation which is refined using morph targets and blend shapes, resulting in a highly realistic facial animation. Game engines and photogrammetry is used to generate a photo-realistic human avatar, to which the base mesh is mapped to generate a photorealistic animated avatar. Virtual humans offer several advantages over real persons, including the ability to customise the persons appearance, voice, accent, language, location, mannerisms etc., making them an ideal solution for global education. The process flow will describe the methods, analysis and interpretations for using AI to generate natural photo-realistic avatars, and the potential contributions to the advancements in engineering education. In conclusion, virtual humans have the potential to revolutionise the way education is delivered in a post-COVID world. By combining AI and CGI, photorealistic virtual human avatars can be created that are highly engaging, customisable, and accessible to students all over the world.