【演講資訊】講題:Key Challenges in Computational Cinematography_講者:Prof. Marc Christie_時間:10/23 (四) 11:00~地點:EC115
- Published on
- Author
- 許詩婷
王昱舜教授邀請University of Rennes 1 France的Prof. Marc Christie於10/23 (四) 11:00~進行演講,以下為演講資訊,歡迎踴躍參加,感謝您。
Time:October 23 (Thu)11.00-12.00
Place:EC 115
Abstract
Computational Cinematography is emerging as a central field at the intersection of computer vision, computer graphics, and artilicial intelligence. Its overarching goal is to assist - and increasingly, to augment -- the creation and transformation of cinematic content through computational means. The field spans a wide spectrum of research topics, from advanced image and video processing, to neural 3D reconstruction, to generative methods capable of producing entire virtual scenes in elther 3D or pivel space.
In this talk, I will first outine the evolution of Computational Cinematography over the past decade, emphasizing how recent breakthroughs in neural and generative representations have begun to redefine the boundaries between capture, simulation, and imagination. I will then present our recent contributions, focusing on diffusion-based techiques for controllable 3D and video generation, aimed at preserving cinematic intent while expanding creative possiblities. Building on these results, I will conclude by identifying the key scienfic and technological challenges that must be addressed to make computational tools not only powerful, but also artistically meaningful for the next generation of visual storytelling.