More about HKUST
A Survey of Talking Head Video Generation
PhD Qualifying Examination Title: "A Survey of Talking Head Video Generation" by Mr. Fating HONG Abstract: Talking head video generation is a pivotal research domain within computer vision, encompassing numerous practical applications such as virtual assistants, video conferencing, and animation. The primary challenge in producing high-quality talking head videos lies in the effective transfer of driving signals from modalities—such as audio or facial expressions—to a reference video or image. This survey provides a comprehensive review of advancements in the field of talking head video generation, with a particular emphasis on deep learning methodologies. We begin by defining the problem and underscoring its significance in contemporary technological contexts. Subsequently, we discuss various existing methods proposed to address this issue, categorizing them into audio-driven and expression-driven approaches. Representative works within each category are critically examined from multiple perspectives, highlighting seminal contributions and analyzing their respective advantages and limitations. Ultimately, this survey aims to serve as a foundational resource for future research endeavors and to stimulate innovative ideas for addressing this challenging problem. Date: Thursday, 19 June 2025 Time: 3:00pm - 5:00pm Venue: Room 3494 Lifts 25/26 Committee Members: Dr. Dan Xu (Supervisor) Dr. Hao Chen (Chairperson) Dr. Qifeng Chen