The article discusses uTalk, an intelligent AI system developed to improve human-computer interaction. uTalk leverages the capabilities of ChatGPT and other technologies, including Whisper, Microsoft Speech Services, and the talking head system SadTalker. This system allows users to converse with an avatar, ask questions, and generate content. The process involves converting the user’s spoken question into text, feeding it into ChatGPT, and generating a response. The response is then converted into speech and combined with an avatar image to create a talking portrait. The system’s performance improved by 9.8% after integrating and parallelizing SadTalker with Streamlit.
Publication date: 5 Oct 2023
Project Page: Not provided
Paper: https://arxiv.org/pdf/2310.02739
