Research led by undergraduate researchers at Jeonbuk National University (JBNU) has been accepted by a world-leading conference, drawing attention.

JBNU said on the 30th that research conducted by undergraduate researchers Bang Minhyuk and Kim Junhyung of the Vision Intelligence Lab, Department of Computer Science & Artificial Intelligence (Advisor: Professor Seho Lee), was accepted to ECCV (European Conference on Computer Vision) 2026, one of the world's three leading computer vision conferences.

ECCV, together with CVPR and ICCV, is one of the world's three leading international computer vision conferences. It is a top-tier academic forum where researchers from universities, research institutions, and global companies present the latest research achievements.

The accepted paper, "Geometry-Aware Style Transfer in 3D Gaussian Splatting," proposes a new style-transfer technique that applies artistic styles to scenes based on 3D Gaussian Splatting (3DGS) while preserving geometric structure and spatial information. Unlike prior methods that were limited to reflecting surface properties such as color or texture, the proposed approach is designed to capture the structural characteristics of the reference style as well.

The research team introduced a training architecture that incorporates the scene's geometric information, successfully applying the style's form and structure while stably preserving the original scene's structure. Through various experiments, they demonstrated superior visual quality and structure-preservation performance compared to existing methods. They also suggested potential applications in fields such as the metaverse, digital content production, virtual reality (VR), and augmented reality (AR).

This achievement is the first case in which undergraduate-led research from the Vision Intelligence Lab at JBNU has had a paper accepted to a top-tier computer vision conference, and is regarded as a significant demonstration of student-centered research capability. Co-first authors Bang Minhyuk and Kim Junhyung are fourth-year students in the Department of Computer Science & Artificial Intelligence, and achieved this result after approximately one year of research.

Professor Seho Lee, who supervised them, said, "ECCV is a world-class conference where even master's and doctoral students find it difficult to have papers accepted. It is highly meaningful that undergraduate researchers led the research and achieved such a result."

He added, "This result stems from the students' challenges and research capabilities, and we will continue to actively support the cultivation of world-class AI talent."

This research was carried out with support from the National Research Foundation of Korea's Individual Basic Research (Early-Career Research) Program.