Simon Fraser University has launched a groundbreaking lab that aims to revolutionize the post-production process for independent Canadian filmmakers through the use of analytical AI tools.
Short Summary:
- SFU’s Computational Photography Lab focuses on solving fundamental computer vision problems.
- The lab is creating CGI-like capabilities for real-world footage.
- Local filmmakers and students are invited to collaborate, impacting AI research direction.
Simon Fraser University (SFU) has embarked on a transformative journey with the inauguration of its new Computational Photography Lab, envisioned to pioneer the development of advanced AI post-production tools. This initiative, headed by Yağız Aksoy, a key figure in the School of Computing Science, underscores SFU’s ambition to bolster the Canadian film industry on a global scale.
Aksoy, an expert in computer vision whose work aims to decode the interaction of light and geometry in images, articulates his vision succinctly: “Our aim is to bring the capabilities of CGI tools such as the Unreal Engine to post-production software for live-action scenes.” The professor’s dedication towards this technological frontier is reflected in the meticulously designed research studio at SFU’s Burnaby campus, equipped with an array of cameras positioned before a vibrant green screen.
“With a physical understanding of real scenes, CGI capabilities such as controlling light sources and the motion of cameras become possible for real-world footage,” Aksoy explains.
The lab’s setup isn’t just an academic playground. Aksoy and his team produce educational videos to serve dual purposes: research material and practical insights into everyday challenges faced by filmmakers. The lab’s vision extends beyond academic walls, aiming to be a collaborative hub for local independent filmmakers and film students who will co-create content while informing AI’s functional trajectory.
“This will help us understand the everyday challenges and the expectations that the creatives have, and also allow creatives to tell AI what to do, which will guide our research direction,” Aksoy says.
The thrust of SFU’s initiative centers on analytical AI, distinguishing itself from more common generative models like ChatGPT or Stable Diffusion. Instead of generating content, analytical AI analyzes existing footage, enabling filmmakers to manipulate light and geometry post-shoot.
Aksoy emphasizes the symbiotic evolution of technology and storytelling within the cinematic realm, positing that AI-driven tools can enhance low-budget productions by offering high-caliber post-production capabilities without inflating costs. The anticipated outcome? A richer visual language and specialized storytelling techniques.
“When you make profits the focus, AI becomes a threat to replace creative jobs. But AI is just a tool,” Aksoy asserts. His work aims to empower filmmakers rather than replace them.
Notably, the lab’s endeavors received substantial support via a $200,000 grant from the Canada Foundation for Innovation’s John R. Evans Leaders Fund, facilitating the acquisition of cutting-edge equipment. This newly equipped studio is poised to refine methodologies such as monocular depth estimation, crucial for accurate scene representation from single images.
“With the hardware we are getting from the CFI JELF grant, we will be able to generate better and higher quality training data which will have a large impact on the applicability of our machine learning models,” Aksoy mentions.
Aksoy’s forward-thinking vision doesn’t begin and end with local talent; it reverberates internationally. His lab’s research, like the acclaimed paper on intrinsic decomposition, potentially redefines image analysis, broadening the horizon for computational photography and interactive AI applications globally.
“I dream of seeing artists create things that I can’t even imagine using the tools that we develop,” says Aksoy.
The overarching aim is to craft AI tools that filmmakers can commandeer with precision, thereby democratizing high-quality post-production across varying budget scales. This mission aligns with broader strategic goals, reflecting Canada’s commitment to nurturing innovative research and practical applications.
Collaborations are integral to the lab’s ethos. By connecting with entities like the Vancouver Film School and local startups, Aksoy ensures a continual feedback loop, enriching both the AI models and the filmmaking protocols. This interaction fosters a dynamic ecosystem where technological advances and creative needs coalesce seamlessly.
The lab’s commitment extends to younger generations of filmmakers through initiatives like The Cinematheque’s educational programs. These programs emphasize comprehensive filmmaking from scripting to post-production, reflecting an inclusive and hands-on educational philosophy.
SFU’s holistic approach promises not just academic advancements but practical tools that reshape the Canadian film industry, allowing it to compete robustly on an international scale. By grounding AI capabilities in real-world applications and continuous feedback, SFU’s lab is poised to set new standards in cinematic excellence.
