Jaesung Choe

Hi, I am a research scientist at NVIDIA Research Taiwan. My main research interest is Multimodal Large Language Model for 3D data.

I received my PhD from KAIST under the supervision of In So Kweon. After the PhD, I worked as a Postdoc at POSTECH under Jaesik Park. I was a research intern at NVIDIA Research AI Algorithm with Christopher Choy, Anima Anandkumar, and at Meta Reality Lab with Shuochen Su.

Email  /  Google Scholar  /  CV  /  LinkedIn  /  Twitter

profile photo
News

If you're interested in a 2025 internship, please (1) send me an email, submit your application at (2) GoogleForm and (3) Workday.

Research

I'm interested in computer vision for 3D scene understanding, such as (1) 3D Visual Language Model, (2) 3D Scene Reconstruction and Neural Rendering and (3) 3D Vehicle/Object Perception. Some papers are highlighted.

[Under review] 3D Visual Language Model
Junha Lee*, Chunghyun Park*, Jaesung Choe, Jonathan Tremblay, De-An Huang, Yu-Chiang Frank Wang, Jan Kautz, Minsu Cho, Christopher Choy
*Equal contribution

Multimodal LLM for 3D scene understanding focusing on the dense captioning and scene captioning tasks.
[Under review] 3D Gaussian Splatting
Cheng Sun, Jaesung Choe, Yu-Chiang Frank Wang

New rasterization method for 3D Gaussian Splatting
Spacetime Surface Regularization for Neural Dynamic Scene Reconstruction
Jaesung Choe, Christopher Choy, Jaesik Park, In So Kweon, Anima Anandkumar
ICCV 2023
Paper

Spacetime surface regularization technique for 4D surface reconstruction and dynamic scene rendering.
PointMixer: MLP-Mixer for Point Cloud Understanding
Jaesung Choe*, Chunghyun Park*, Francois Rameau, Jaesik Park, In So Kweon
*Equal contribution
ECCV 2022
Paper / Code / Poster / Video

Design a new MLP-only architecture for 3D points.
Deep Point Cloud Reconstruction
Jaesung Choe, ByeongIn Joung, Francois Rameau, Jaesik Park, In So Kweon
ICLR 2022
Paper

Perform point cloud upsampling and denoising tasks simultaneously. Demonstrate good generalization performance.
Facial Depth and Normal Estimation using Single Dual-Pixel Camera
Minjun Kang, Jaesung Choe, Hyowon Ha, Hae-Gon Jeon, Sunghoon Im, In So Kweon, Kuk-Jin Yoon
ECCV 2022
Paper / Code

Face reconstruction method using a monocular Dual-Pixel camera.
VolumeFusion: Deep Depth Fusion for 3D Scene Reconstruction
Jaesung Choe, Sunghoon Im, Francois Rameau, Minjun Kang, In So Kweon
ICCV 2021
Paper / Video / Slide

Deep learning based depth fusion algorithm for indoor scene reconstruction.
Volumetric Propagation Network: Stereo-LiDAR Fusion for Long-Range Depth Estimation
Jaesung Choe, Kyungdon Joo, Tooba Imtiaz, In So Kweon
RA-L 2021
Paper (RA-L) / Paper (ICRA) / Video

Depth estimation using LiDAR pointcloud and stereo images.
Mentorship

Yoonwoo Jeong (research intern - 2024)
Chunghyun Park (research intern - 2023)

Academic collaboration

Ji-Yeon Kim (POSTECH - 2024)
Jun-Seong Kim (POSTECH - 2024)
Minjun Kang (KAIST - 2023)

Talks

Pusan Univ., Sep 2024
Hangyang Univ., Sep 2024
DGIST, April 2024
Kakao Brain, May 2022


Source code for this page was taken from Jon Barron's website.