HTTP/2 200 server: GitHub.com content-type: text/html; charset=utf-8 last-modified: Fri, 26 Dec 2025 07:16:13 GMT access-control-allow-origin: * strict-transport-security: max-age=31556952 etag: W/"694e363d-3442" expires: Sun, 28 Dec 2025 07:30:52 GMT cache-control: max-age=600 content-encoding: gzip x-proxy-cache: MISS x-github-request-id: F989:2685F2:7492E1:82BE4A:6950DA54 accept-ranges: bytes age: 0 date: Sun, 28 Dec 2025 07:20:52 GMT via: 1.1 varnish x-served-by: cache-bom-vanm7210054-BOM x-cache: MISS x-cache-hits: 0 x-timer: S1766906453.762767,VS0,VE200 vary: Accept-Encoding x-fastly-request-id: 7526868a74caf7deef10333a7d43d8b514992145 content-length: 3501

Taeksoo Kim Hi, I'm a 3rd year MS/PhD student at Seoul National University, where I am advised by Prof. Hanbyul Joo. My current research interests lie in simulating the physical world with humans using video generative models. I received a B.S in Industrial Engineering and a B.S. in Computer Science and Engineering from Seoul National University. Email  /  CV  /  Scholar  /  Github News Feb. 2024: Our work, GALA: Generating Animatable Layered Assets from a Single Scan, got accepted in CVPR 2024! Jul. 2023: Our work, NCHO: Unsupervised Learning for Neural 3D Composition of Humans and Objects, got accepted in ICCV 2023! Mar. 2023: Joined SNU Visual Computing Lab! Research Dexterous World Models Byungjun Kim*, Taeksoo Kim*, Junyoung Lee, Hanbyul Joo arXiv, 2025 project page / code / arXiv We present DWM, a scene-action-conditioned video diffusion model that simulates dexterous human interactions in static 3D scenes. Target-Aware Video Diffusion Models Taeksoo Kim, Hanbyul Joo arXiv, 2025 project page / code & data / arXiv Given an input image, our target-aware video diffusion model generates a video in which an actor accurately interacts with the target, specified with its segmentation mask. GALA: Generating Animatable Layered Assets from a Single Scan Taeksoo Kim*, Byungjun Kim*, Shunsuke Saito, Hanbyul Joo CVPR, 2024 project page / code / arXiv We present a framework that takes as input a single-layer clothed 3D human mesh and decomposes it into complete multi-layered 3D assets. NCHO: Unsupervised Learning for Neural 3D Composition of Humans and Objects Taeksoo Kim, Shunsuke Saito, Hanbyul Joo ICCV, 2023 project page / code & data / arXiv We present a novel framework for learning a compositional generative model of humans and objects (backpacks, coats, scarves, and more) from real-world 3D scans.

Template from Jon Barron's website