Portrait of Mijeong Kim

Hello, I'm Mijeong Kim

Ph.D. Candidate · Computer Vision Lab, SNU

My research focuses on 3D/4D reconstruction and world modeling. I develop probabilistic representations to reconstruct dynamic scenes from limited observations, with recent work extending to multi-view attention and text-controllable physics simulation.

My undergraduate studies in naval architecture and ocean engineering at SNU gave me a foundation in robotics, mechanics, and computational fluid dynamics (CFD), which now informs my computer vision research on dynamic 3D scenes.

Looking for full-time opportunities. I am completing my Ph.D. and seeking Research Scientist / Applied Scientist roles starting from December 2026. See my CV

Research Overview

Ongoing

Probabilistic World Modeling & Text-Controllable Physics

  • arXiv:2610.163371st author — WorldRoPE: world-centric attention across views
  • arXiv:2610.306791st author — AgentDreamer: text-controllable 3D motion and physics

2025–26

Physics-Aware Vision & Probabilistic 4D Reconstruction

  • CVPR 20261st author — PhysGaia: physics-aware multi-body benchmark
  • CVPR 20261st author — GP-4DGS: probabilistic 4D reconstruction from monocular video
  • WACV 20261st author — HyperPose: implicit pose embeddings for 3D generation

2024

Uncertainty-Aware 4D Gaussian Splatting

  • NeurIPS 20241st author — UA-4DGS: uncertainty-aware 4D reconstruction from monocular video

2023

3D Scene Generation with Neural Fields

  • NeurIPS 2023 — 3D scene generation through mixtures of neural fields

2022

Few-Shot NeRF & Image Alignment

  • CVPR 20221st author (300+ citations) — InfoNeRF: few-shot rendering via ray-entropy minimization
  • MVA 20221st author — Nonparametric image alignment via graph convolutions

News

Publications

WorldRoPE: Probabilistic World-Centric Attention with 2D Patch Tokens

WorldRoPE: Probabilistic World-Centric Rotary Embedding for Multi-View Attention

Mijeong Kim, Kyujin Lim, Bohyung Han

arXiv:2610.16337 (Under Review)

AgentDreamer: Agentic 3D Motion Reasoning for Text-Controllable Physics Simulation — research illustration

AgentDreamer: Agentic 3D Motion Reasoning for Text-Controllable Physics Simulation

Mijeong Kim*, Suho Park*, Junha Park, Bohyung Han

arXiv:2610.30679 (Under Review)

GP-4DGS: Probabilistic 4D Gaussian Splatting from Monocular Video via Variational Gaussian Processes — research illustration

GP-4DGS: Probabilistic 4D Gaussian Splatting from Monocular Video via Variational Gaussian Processes

Mijeong Kim, Jungtaek Kim, Bohyung Han

CVPR 2026

PhysGaia: A Physics-Aware Benchmark with Multi-Body Interactions for Dynamic Novel View Synthesis — research illustration

PhysGaia: A Physics-Aware Benchmark with Multi-Body Interactions for Dynamic Novel View Synthesis

Mijeong Kim*, Gunhee Kim*, Jungyoon Choi, Wonjae Roh, Bohyung Han

CVPR 2026

HyperPose: Hyper-pose Embeddings for 3D-Aware Generative Models with Self-Supervised Disentangling of Pose and Scene — research illustration

HyperPose: Hyper-pose Embeddings for 3D-Aware Generative Models with Self-Supervised Disentangling of Pose and Scene

Mijeong Kim, Namgi Kim, Bohyung Han

WACV 2026

UA-4DGS: 4D Gaussian Splatting in the Wild with Uncertainty-Aware Regularization — research illustration

UA-4DGS: 4D Gaussian Splatting in the Wild with Uncertainty-Aware Regularization

Mijeong Kim, Jongwoo Lim, Bohyung Han

NeurIPS 2024

Generative Neural Fields by Mixtures of Neural Implicit Functions — research illustration

Generative Neural Fields by Mixtures of Neural Implicit Functions

Tackgeun You, Mijeong Kim, Jungtaek Kim, Bohyung Han

NeurIPS 2023

ContraNeRF: 3D-Aware Generative Model via Contrastive Learning with Unsupervised Implicit Pose Embedding — research illustration

ContraNeRF: 3D-Aware Generative Model via Contrastive Learning with Unsupervised Implicit Pose Embedding

Mijeong Kim, Hyunjoon Lee, Bohyung Han

arXiv:2304.14005

InfoNeRF: Ray Entropy Minimization for Few-Shot Neural Volume Rendering — research illustration

InfoNeRF: Ray Entropy Minimization for Few-Shot Neural Volume Rendering

Mijeong Kim, Seonguk Seo, Bohyung Han

CVPR 2022 (citation 300+)

Beyond homography: nonparametric image alignment via graph convolutional networks — research illustration

Beyond homography: nonparametric image alignment via graph convolutional networks

Mijeong Kim, Sanghyeok Chu, Bohyung Han

MVA (Machine Vision Applications) 2022

Overcoming Forgetting in Federated Learning via Importance From Agent — research illustration

Overcoming Forgetting in Federated Learning via Importance From Agent

Mijeong Kim*, Jiho Kim*, Minji Kim*

Conducted in Machine Intelligence and Data Science Lab.
(430.658 of SNU, Prof. Taesup Moon)

Experience

Huawei · Noah’s Ark Labs

Sep 2025 – Feb 2026

Applied Scientist Intern · London, UK

Feed-forward 4D reconstruction from monocular video.

Kakao Brain

Jul 2022 – Sep 2022

Research Intern · Seongnam, Korea

3D-aware generation via contrastive learning · ContraNeRF

Background & Education

My undergraduate training in robotics, mechanics, and CFD at SNU now informs my computer vision research on dynamic 3D scenes.

Honors & Awards

Personal Life

Outside of research, I enjoy working out and hiking. I sometimes go on night hikes with friends to enjoy the city lights. During my internship in the UK, I went hiking in Sevenoaks with a fellow intern, and the views were beautiful!