Hwan Heo

Hello, I'm AI Researcher

About

Hi, it's a delightful serendipity to meet you in the vast expanse of the digital cosmos. I'm Hwan Heo, a lifelong learner, and explorer in AI & Computer Vision.

Vision AI Researcher

Specializing in 3D Computer Vision.

  • Interest: 3D Vision, Neural Rendering
  • Degree: Master's in Computer Science
  • E-mail: gjghks950@naver.com
  • City: Sungnam, South Korea

Hwan Heo is an AI researcher at Graphics AI Lab, NCSOFT. He holds a B.S. and M.S. in Computer Science from Korea University, under the supervision of Professor Hyunwoo J. Kim. He was Research Intern in Generation Research at NAVER AI Lab mentored by Dr. Jin-Hwa Kim. His research emphasizes 3D Computer Vision, Radiance Fields, and 3D Generation, contributing to the cutting-edge developments in these fascinating fields.

Resume

Download CV

Summary

Hwan Heo

Driven AI Researcher with 3+ years of experience specializing in 3D Computer Vision, focused on developing leading-edge techniques for Neural Rendering and 3D Generation.

  • AI Researcher at NC Research
  • Research Personnel, Alternative Military Service

Education

Master of Computer Science & Engineering

2020 - 2022

MLV Lab, Korea University, Seoul

Bachelor of Computer Science & Engineering

2015 - 2020

Korea University, Seoul

Professional Experience

AI Researcher

2023 - Present

Neural Rendering & 3D Generation, NCSOFT Research

  • Lead in the research on 3D asset generation; 3D Latent Diffusion & Texture Synthesis
  • Application on Neural Rendering; Gaussian Splatting in Game Engine , Fast Neural Facial Avatar for Multi-Camera System

AI Researcher, Intern

2022

Generation Research, NAVER AI Lab

  • Research on Neural Rendering; Joint Optimization of Camera Pose and Hash Grid Radiance Fields

  • Image-to-Image Translation on Equirectangular Images

Portfolio

Research & Development: Featured Publication & Project Lists

  • All
  • Research
  • Application
  • Personal
CaPa teaser
Featured on HuggingFace Papers, Github (150+ stars)
CaPa: Carve-n-Paint Synthesis for Efficient 4K Textured Mesh Generation
3D asset generation scheme that generates 4K Textured Mesh under 30 seconds
NC Research
Pose teaser
ICML 2023
Robust Camera Pose Refinement for Multi-Resolution Hash Encoding
Joint optimization of camera poses and hash-grid NeRF (Instant-NGP)
Naver AI Lab
Occ-NeRF teaser
arXiv 2023
Semantic-aware Occlusion Filtering Neural Radiance Fields in the Wild
Neural scene reconstruction from a small number of unconstrained in-the-wild photos.
Korea University
NeRF Game teaser
Neural Rendering in Game Engine
Integration Gaussian Splatting from 360 cameras to the Game Engine
NC Research
NeRF Face teaser
Featured on Github (210+ stars)
Radiance Fields from Deep-based Structure-from-Motion
Exploring Gaussian Splatting reconstruction from VGGSfM and MaST3R
Personal Project
NeRF light teaser
Featured on Github (130+ stars)
Custom Web Viewer for 2D Gaussian Splatting
First public viewer for 2D Gaussian Splatting using nerfstudio's viser
Personal Project
NeRF light teaser
Specular-Diffuse Disentanglement w/ Hash-Grid NeRF (feat. nerfstudio's NeRFacto)
Light disentanglement modeling on the Instnat-NGP-based NeRF model
NC Research
NeRF Face teaser
Fast Neural Facial Avatar for Multi-Camera System
Fast NeRF face avatar reconstruction from multi-camera system
NC Research
NeRF light teaser
CVPR 2022
Consistency Learning via Decoding Path Augmentation for Transformers in Human Object Interaction Detection
Enhance DeTR-based Human-object interaction detection using consistency regularization
Korea University
NeRF light teaser
IEEE Access 2021
Search-and-Attack: Temporally Sparse Adversarial Attack on Videos
Adversarial Attack for video action recognition model
Korea University


Publications

  • CaPa: Carve-n-Paint Synthesis for Efficient 4K Textured Mesh Generation

    Hwan Heo, Jangyeong Kim, Seongyeong Lee, Jeong-A Wi, Junyoung Choi, Sangjun Ahn
    Technical Report, Featured on HuggingFace's Daily Papers, TLDR AI

  • DreamSurfels: Generative Gaussian Surfels for Geometrically Accurate 3D Content Creation

    Hwan Heo, Sangjun Ahn
    The 15th International Conference on ICT Convergence (ICTC), Oct 2024
    Oral presentation, domestic conference

  • Robust Camera Pose Refinement for Multi-Resolution Hash Encoding

    Hwan Heo, Taekyung Kim, J. Lee, Jiyoung Lee, Soohyun Kim, Hyunwoo J. Kim, Jin-Hwa Kim
    The 40th International Conference on Machine Learning (ICML), Jul 2023

  • Semantic-aware Occlusion Filtering Neural Radiance Fields in the Wild

    Jaewon Lee, Injae Kim, Hwan Heo, Hyunwoo J. Kim
    preprint, arXiv 2023

  • Consistency Learning via Decoding Path Augmentation for Transformers in Human Object Interaction Detection

    Jihwan Park, Seunjun Lee, Hwan Heo, Hyunggyu Choi, Hyunwoo J. Kim
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun 2022

  • Domain Generalization Emerges from Dreaming

    Hwan Heo, Younjin Oh, Jaewon Lee, Hyunwoo J. Kim
    preprint, arXiv 2023

  • Temporally Sparse Adversarial Perturbations on Videos

    Hwan Heo, Dohwan Ko, Jaewon Lee, Youngjoon Hong, Hyunwoo J. Kim
    IEEE Access, Oct 2021 vol. 9, pp. 146938,
    Internatrional journal publication

Blog: link | velog (kor)

Featured Recent Posts:

3D Model Viewer in Web

Deep Dive into 3D Latent Diffusion

NeRF for Tomography (X-Ray) & Quick Viewer Development Tip for Any Radiance Fields

Review & Analysis: Gaussian Ray Tracing

Analysis of Gaussian Splatting Rasterization Algorithm