Yonwoo Choi
yhugestar at gmail dot com

I am currently fulfilling my alternative military service as an Algorithm Engineer at GenGenAI, where I focus on generating synthetic data with 3D.

My research focuses on the intersection of 3D computer vision and generative models, developing methods to create realistic synthetic data for downstream tasks.

I completed my M.S at Seoul National University and my B.A. at the University of Virginia both in computer science. During my M.S, I was fortunate to be part of the Visual Computing Lab under the supervision of Hanbyul Joo. Previously, I have worked at SECERN AI, working on Animatable 3D Human Avatars.

CV  /  Google Scholar  /  OpenReview  /  GitHub  /  Linkedin

News
  • 2025.09 I started working as an Algorithm Engineer at GenGenAI.
  • 2025.03 SVAD has been accepted to CVPR 2025 Workshop SyntaGen.
  • 2024.03 I started working as an Research Engineer at SECERN AI.
  • 2024.02 I've graduated from SNU with an M.S.
Papers
SVAD: From Single Image to 3D Avatar via Synthetic Data Generation with Video Diffusion and Data Augmentation
Yonwoo Choi
CVPR 2025 Workshop
SyntaGen-Harnessing Generative Models for Synthetic Visual Datasets
, 2025
project page / paper / code / arXiv / review

3D Animatable Avatar Generation from a Single Image

A Wrist Mounted Camera System for Hand Pose Estimation from Single RGB Images
Yonwoo Choi, Hanbyul Joo
Masters Thesis
Computer Science and Engineering, Seoul National University
, 2024
paper

A System for Collecting Pseudo Ground Truth 3D Hand Poses to Improve Pose Estimation Models

Opensource
Single Image to 3D Human Avatar
SECERN AI
May 2025
Code GitHub stars
- Official Implementation of the paper SVAD
- Synthetic Data Generation Pipeline for 3DGS Avatar Training
Facebuilder
SECERN AI
March 2025
Code GitHub stars
- 3D Face Mesh Deformation with Pin-Based Interaction
- Multiview Face Modeling with Real-time 3D Consistency
SMPL Fitting
SECERN AI
June 2024
Code GitHub stars
- Fitting SMPL Parameters to 3D Human Joints
- Parametric Human Shape and Pose Optimization
Egocentric Human Pose Estimator
Visual Computing Lab
December 2023
Code GitHub stars
- Unofficial Implementation of the Paper xR-EgoPose
- Egocentric 3D Human Pose Estimation from a Single Fisheye Camera
Multiview Human Pose Capture System
Visual Computing Lab
April 2022
Code GitHub stars
- 5 GoPro Camera Setup for 3D Human Reconstruction
- Camera Calibration with Zhang's Method
Work Experience
GenGenAI
Algorithm Engineer
September 2025 - Present
- Generating Synthetic Data with 3D
SECERN AI
Research Engineer
March 2024 - September 2025
- Animatable 3D Human Avatars
Teaching
Teaching Assistant, Seoul National University Fall 2023
4190.408 001 - Artificial Intelligence (Instructor: Hanbyul Joo)
Teaching Assistant, Seoul National University Spring 2022
4190.773 002 - 3D Computer Vision (Instructor: Hanbyul Joo)