profile photo

Sihun Cha

/sihʌn tɕʰɑ/

Email: chacorp@kaist.ac.kr

I am a Ph.D. student at Visual Media Lab, KAIST, advised by Prof. Junyong Noh. I studied Fine Arts as a Bachelor at Korea National University of Arts. I envision techniques for real-world applications and bring these ideas to daily life through implementation in various ways. My research has primarily focused on facial animation. During my Ph.D., I came to recognize that the stability and reliability of models are closely tied to the representations they learn. Recently, I have been exploring directions centered on representation.

Scholar Github

Eduation

한국과학기술원 | Korea Advanced Institute of Science and Technology
Ph.D. student in Computer Science, advised by Junyong Noh.
2022 - Present
한국과학기술원 | Korea Advanced Institute of Science and Technology
M.S. in Computer Science, advised by Junyong Noh.
2020 - 2022
한국예술종합학교 | Korea National University of Arts
B.F.A. in Fine Arts
2013 - 2020

Experience

CLO Virtual Fashion Inc.
Machine Learning Engineer
Garment generation and manipulation.
Apr 2026 - Present
Seoul, South Korea
Visual Media Lab, KAIST
Research Assistant.
Topics of texture generation, 3D facial animation.
Jan 2021 - Present
DaeJeon, South Korea
Flawless AI
Research Scientist Intern
Audio to Mouth Synthesis using Diffusion Model.
Sep 2024 - Dec 2024
Santa Monica, US

Publications

One-shot Motion Personalization for Audio-driven Portrait Video Generation

One-shot Motion Personalization for Audio-driven Portrait Video Generation

Sihun Cha, Kwanggyoon Seo, Akin Caliskan, Pablo Garrido, Hyeongwoo Kim, Bindita Chaudhuri, Junyong Noh
In submission

Finetuning audio-driven portrait video diffusion model using a single viewo for personalized motion

X-AVDT: Audio-Visual Cross-Attention for Robust Deepfake Detection

X-AVDT: Audio-Visual Cross-Attention for Robust Deepfake Detection

Youngseo Kim, Kwan Yun, Seokhyeon Hong, Sihun Cha, Colette Suhjung Koo, Junyong Noh
CVPR 2026

Robust deepfake detection method and cross-generator dataset by leveraging cross-attention features to capture audio-visual correlations.

Paper Project Page Code
Mesh Agnostic Audio-Driven 3D Facial Animation

Mesh Agnostic Audio-Driven 3D Facial Animation

* Kwanggyoon Seo, * Sihun Cha, Hyeonho Na, Inyup Lee, Junyong Noh (*equal contribution)
KCGS 2025 (Best paper)

An end-to-end method for animating a 3D face mesh with arbitrary shape and triangulation from a given speech audio.

Paper
Neural Face Skinning for Mesh-agnostic Facial Expression Cloning

Neural Face Skinning for Mesh-agnostic Facial Expression Cloning

Sihun Cha, Serin Yoon, Kwanggyoon Seo, Junyong Noh
Eurographics 2025

A method that enables direct retargeting between two facial meshes with different shapes and mesh structures.

Paper Project Page Code
Deep Learning-Based Facial Retargeting Using Local Patches

Deep Learning-Based Facial Retargeting Using Local Patches

* Yeonsoo Choi, * Inyup Lee, Sihun Cha, Seonghyeon Kim, Sunjin Jung, Junyong Noh (*equal contribution)
Eurographics 2025

Retargeting facial expression from a source human performance video to a target stylized 3D character using local patches.

Paper
NeRFFaceSpeech: One-shot Audio-diven 3D Talking Head Synthesis via Generative Prior

NeRFFaceSpeech: One-shot Audio-diven 3D Talking Head Synthesis via Generative Prior

Gihoon Kim, Kwanggyoon Seo, Sihun Cha, Junyong Noh
CVPRW 2024

One-Shot Audio-driven 3D talking head generation with enhanced 3D consistency using NeRF and generative knowledge from single image input.

arXiv Project Page Code
LeGO: Leveraging a Surface Deformation Network for Animatable Stylized Face Generation with One Example

LeGO: Leveraging a Surface Deformation Network for Animatable Stylized Face Generation with One Example

* Soyeon Yoon, * Kwan Yun, Kwanggyoon Seo, Sihun Cha, Jung Eun Yoo, Junyong Noh (*equal contribution)
CVPR 2024

Creating an animatable stylized 3D face mesh with one example pair.

Paper arXiv
Project Page Code
Generating Texture for 3D Human Avatar from a Single Image using Sampling and Refinement Networks

Generating Texture for 3D Human Avatar from a Single Image using Sampling and Refinement Networks

Sihun Cha, Kwanggyoon Seo, Amirsaman Ashtari, Junyong Noh
Eurographics 2023, Computer Graphics Forum (CGF)

Generating 3D human texture from a single image using sampling and refinement process by utilizing geometry information.

Paper arXiv
Project Page Code
Reference Based Sketch Extraction via Attention Mechanism

Reference Based Sketch Extraction via Attention Mechanism

Amirsaman Ashtari, Chang Wook Seo, Cholmin Kang, Sihun Cha, Junyong Noh
SIGGRAPH Asia 2022, ACM Transactions on Graphics (TOG)

Extracting a sketch from an image in the style of a given reference sketch while preserving the visual content of the image.

Paper Code
Generating 3D Human Texture from a Single Image with Sampling and Refinement

Generating 3D Human Texture from a Single Image with Sampling and Refinement

Sihun Cha, Kwanggyoon Seo, Amirsaman Ashtari, Junyong Noh
SIGGRAPH 2022 Posters

A method for generating 3D human texture from a single image based on SMPL model, using sampling and refinement process.

Paper Code
“Anyway,”: Two-player Defense Game via Voice Conversation

“Anyway,”: Two-player Defense Game via Voice Conversation

* Minki Hong, * YoungJun Choi, * Sihun Cha (*equal contribution)
CHI Play 2021

A two-player conversational defense game that uses voice conversation as an input.

Paper

Project

Development of Universal Fashion Creation Platform Technology for Avatar Personality Expression

Development of Universal Fashion Creation Platform Technology for Avatar Personality Expression

as a Project Manager (Intra)
Jun. 2023 ~ Ongoing

Development of universal fashion creation technology and creative platform that enables general users self-expression through intuitive avatar creation

Development of Core SW Technology for Realistic 3D Facial Animation Generation

Development of Core SW Technology for Realistic 3D Facial Animation Generation

as a Project Manager (Intra)
Aug. 2022 ~ Jan. 2024

Developing the AI model for 3D facial animation and research on motion retargeting techniques between human and character

3D Cinemagraph for AR Contents Creation

3D Cinemagraph for AR Contents Creation

as a Software Developer
Jun. 2021 ~ Dec. 2022

Development of user-friendly content production technology that enables general users to easily transform a single image into immersive AR content where background and characters within the image move and interact with real-world objects.

Development of self-evolving AI Creation Platform

Development of self-evolving AI Creation Platform

as a Software Developer
Jun. 2021 ~ Dec. 2022

Development of user-friendly animation creation platform through analysis of user input keywords and images for single creators