Sejong University's research institute accelerating development of next-generation GUP technology
Professor Park Woo-chan, center, director of the Future GPU Research Institute at Sejong University, poses with researchers to mark the institute’s opening in May. Courtesy of Sejong University
New technology could be applied to autonomous vehicles, medical imaging
The Future GPU Research Institute of Sejong University is accelerating the development of core technologies for next-generation graphics processing units (GPU) and artificial intelligence (AI) semiconductors, the university said Monday.
The institute, led by professor Park Woo-chan of the department of computer sciences and engineering, is focusing on developing technologies for real-time path tracing, the next frontier in GPU rendering beyond conventional ray tracing.
Ray tracing is a computer graphics technique that renders highly realistic images by simulating how light travels and interacts with objects in a scene. Path tracing is an advanced form of ray tracing that simulates the complete path of light as it bounces around a scene, producing images that are extremely close to real-world lighting.
While path tracing enables ultrarealistic graphics that are virtually indistinguishable from reality, the severe noise generated during real-time rendering has long remained the biggest hurdle to its commercialization.
To address these challenges, the institute is developing a proprietary real-time AI denoising semiconductor technology. Through years of dedicated research, Park’s team has built a strong technological foundation, filing and securing dozens of patents that underscore its competitive advantage in the field.
Unlike conventional GPUs and neural processing units (NPUs), the institute’s AI semiconductor technology guarantees real-time processing while performing AI computations without accessing external memory, enabling an ultralightweight, highly efficient hardware implementation.
The university said the technology has broad potential applications across industries, including surveillance cameras, autonomous vehicles, medical imaging and defense.
The institute, affiliated with the university’s Industry-Academy Cooperation Foundation, has expanded its research to audio AI semiconductor technology for applications such as hearing aids and AI-powered smart glasses.
It is expanding collaborations with leading research centers and industry partners at home and abroad while actively promoting the commercialization of its patented technologies.
“Next-generation GPU technology is a highly advanced field in which global leaders such as Nvidia, AMD and Intel continue to make substantial investments in research and development,” Park said. “We remain committed to pushing the boundaries of innovation to produce world-class research.”
He said the institute will accelerate the successful commercialization of its technologies to emerge as a global leader in next-generation GPUs and AI semiconductors.
Park is an internationally recognized expert in ray tracing GPU technology and has been a leading figure in GPU research in Korea. His research has contributed to advances in ray tracing, a core graphics technology used in gaming, filmmaking and virtual reality.
As demand for photorealistic graphics continues to grow, competition in the field has intensified, with most high-performance and mobile GPUs released by global technology leaders — including Nvidia, AMD, Intel, Apple, Qualcomm and Samsung Electronics — now supporting real-time ray tracing.
Meanwhile, Park is conducting two research projects funded by the Ministry of Science and ICT.
The first project focuses on ultrarealistic extended reality technologies for a metaverse that seamlessly connects the real and virtual worlds. The second one is aimed at advancing intellectual property and commercializing real-time AI denoising hardware technology.