2026.08
Two papers from our collaborations have been accepted to the European Conference on Computer Vision (ECCV) 2026.
Extending a Large View Synthesis Model for Multi-view Panoptic Segmentation
With Kwonyoung Ryu, In-Jae Lee, Jonghyun Jin, Hyunjee Lee, and Jaesik Park. The work extends a large view synthesis model so that a single feed-forward pass produces multi-view-consistent panoptic segmentation together with novel-view synthesis, bridging 3D reconstruction and scene understanding.
URHead: A Unified UV-Space Representation for Joint Mesh–3DGS Optimization in Head Avatars
With Seonghak Lee, Junhee Cho, Jisoo Park, Min-Gyu Park, Ju Hong Yoon, and Junseok Kwon. URHead introduces a unified UV-space representation that jointly optimizes a mesh and 3D Gaussian Splatting, enabling high-fidelity head avatars with consistent geometry and appearance.