Jaehyeok Shim (재혁 심)

Education

Ulsan National Institute for Science and Technology (UNIST)
Ulsan, Republic of Korea
Researcher in Vision-3D Lab, supervised by professor Kyungdon Joo.

Lab Homepage
Sep. 2023 ~ Present
Ulsan National Institute for Science and Technology (UNIST)
Ulsan, Republic of Korea
Master's student in Vision-3D Lab, supervised by professor Kyungdon Joo.

Lab Homepage
Sep. 2021 ~ Aug. 2023
Seoul National University for Science and Technology (Seoultech)
Seoul, Republic of Korea
Undergraduate student, Information and Electricity Engineearing.
Undergraduate researcher in Visual Computing Lab, supervised by professor Yeejin Lee.

Lab Homepage
Mar. 2015 ~ Aug. 2021
Mar. 2020 ~ Aug. 2021

Publications

Jaehyeok Shim, Kyungdon Joo

DITTO addresses a crucial task in 3D computer vision: implicit 3D reconstruction from point clouds. DITTO has significantly enhanced 3D understanding ability by leveraging latent features in two modalities: grid and point latents. The core innovation lies in exploiting the complementary synergy between these two types of latents. Our contribution is effectively integration of these two latent types within a network architecture, improving 3D reconstruction performance. Our research holds significant potential for various applications involving implicit fields.

Project Page   |   Paper   |   Code

Contactgen introduces a novel approach for generating 3D human poses that interact realistically with a given another human. We utilize a guided diffusion framework, optimizing human poses to ensure physically plausible interactions. This optimization is based on the predicted contact area determined by the given type of interaction.

Project Page   |   Paper   |   Code

Jaehyeok Shim, Changwoo Kang, Kyungdon Joo

SDF-Diffusion is a 3D shape generation framework using diffusion models and signed distance fields for continuous 3D representations (such as meshes), offering high-resolution shapes in a two-stage process (generation and super-resolution) and demonstrating competitive performance on unconditional and conditional 3D shape generation tasks.

Project Page   |   Paper   |   Code

Open Source Contributions

kitsu
Code stack of Pytorch boilerplate codes including a DDP-based trainer similar to pytorch-lightning.

space-filling-pytorch
Library for Space Filling Curve (e.g., Hilbert-Curve, Z-Order) implementations based on Triton.

fast-GeM
Generalized Mean Pooling (GeM) implementation using Triton.

GEGLU-triton
Triton implementation of GEGLU.

AI Challenges