CoolMoves

Current Virtual Reality (VR) systems are bereft of stylization and embellishment the user's motion - concepts that have been well explored in animations for games movies. We present CooIMoves, a system expressive accentuated full-body synthesis virtual avatar real-time, from limited input cues afforded by current consumer-grade VR systems, specifically headset hand positions. make use existing capture databases as template repository to draw from. match similar spatio-temporal motions database then interpolate between them using weighted distance metric. Joint prediction probability is used temporally smooth synthesized motion, human dynamics prior. This allows our work even with very sparse (e.g., only 3-5 per action). validate four experiments: technical evaluation quantitative pose reconstruction three additional user studies evaluate quality, embodiment agency.