Motion Capture, Synthesis and Perception for Creating Realistic and Intelligent Virtual Characters
Character animation plays a key role in Games, Human-Computer Interaction (HCI), Virtual Reality (VR), Augmented Reality (AR) applications and Robotics simulations, In this talk, I will present my work that uses motion capture, synthesis and perception methods to create realistic and intelligent virtual characters. First, I will introduce approaches for synthesizing conversational gestures well-coordinated with body motion and synchronized with spoken text and prosody emphasis. I will also give an overview of several hand motion capture and tracking projects for adding detailed finger motions to characters. Furthermore, I will demonstrate motion synthesis approaches that use Machine Learning to infuse intelligence into virtual characters.
Yingying Wang obtained her Ph.D. in Computer Science from University of California, Davis in 2017. Her research interests are Character Animation, Vision and Machine Learning. She currently works at Snap as a researcher and developer. Previously, she also worked at Disney Research and Autodesk. Her research work includes gesture synthesis, hand motion capture, motion style transfer, motion retrieval, 3D human pose estimation and related innovations. She has 13 papers published at peer-reviewed conferences and journals, and won the best paper award from Motion in Games, 2015. She also has five patents filed during her work at Disney Research and Snap Research. She is dedicated to bringing more intelligence to virtual characters and is constantly contributing to several open source projects.