Presented by Ilke Demir, Senior Research Scientist at Intel
One picture is worth a thousand words. What about 100 simultaneous videos to reconstruct every frame of life in a dome? Is this data enough to reconstruct and digitize us realistically? Similar to other industries, the entertainment industry is also being reshaped by AI, especially towards AR/VR consumption. Before democratization of AI and data, such immersive experiences were lacking an essential element: photorealism. As our data increased, our models got deeper, and the reality became decipherable.
This talk will introduce recent deep learning advancements in 3D vision, reconstruction, and shape understanding techniques with a focus on generative models to digitize performances and scenes. In addition to a discussion about how to process such large visual data, the challenges that come with scaling 10x over current capture platforms, and over 200x over state-of-the-art datasets will be presented. The talk will conclude with a sneak peek of upcoming productions in the world’s largest volumetric capture stage at Intel Studios.
00:00 Intro
4:15 The World's Largest Volumetric Stage
7:11 History
8:32 Challenges
23:22 Results
27:35 Conclusion
