The goal of this project is simplify avatar creation by using smartphone data with RGBD sensors, eliminating the need for complex camera arrays shown in Fig 1. This approach enhances scalability and adoption of avatars, as illustrated in Fig 2.
Thus, our problem statement boils down to,
Real-time and Online reconstruction of dynamic scenes with non-rigid motion from monocular RGB-D videos (e.g. from an iPhone) for flexible and interactive engagement.