Byeongjoo Ahn (안병주) | Research Scientist at Apple

I am a research scientist at Apple in Seattle, working on image and video generation foundation models for Apple Intelligence.

I received my PhD from Carnegie Mellon University, where I was advised by Aswin C. Sankaranarayanan and Ioannis Gkioulekas. I earned BS and MS degrees from Seoul National University, where I worked with Kyoung Mu Lee. I also worked as a research scientist at the Korea Institute of Science and Technology.

My first name is pronounced as “Be-Young-Joo.”

Professional Activities

Senior Area Chair: ICML (2027)
Area Chair: CVPR (2026, 2027), ICLR (2027), NeurIPS (2024, 2025 (Top Area Chair), 2026 (Top Area Chair)), ICML (2025)
Reviewer: CVPR, ICCV, ECCV, BMVC, ICLR, NeurIPS, SIGGRAPH, SIGGRAPH Asia, ICCP, TIP, TCI

Research

AToken: A Unified Tokenizer for Vision

Jiasen Lu^*, Liangchen Song^*, Mingze Xu, Byeongjoo Ahn, Yanjun Wang, Chen Chen, Afshin Dehghan, Yinfei Yang

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026 (Oral)

project page · arXiv · code

A unified visual tokenizer for high-fidelity reconstruction and semantic understanding across images, videos, and 3D assets.
Novel-view Acoustic Synthesis from 3D Reconstructed Rooms

Byeongjoo Ahn, Karren Yang, Brian Hamilton, Jonathan Sheaffer, Anurag Ranjan, Miguel Sarabia, Oncel Tuzel, Jen-Hao Rick Chang

Interspeech, 2024

project page · arXiv · video · code · headset recommended

Estimating spatial sound in reconstructed scenes by combining blind audio recordings with 3D information.
Neural Kaleidoscopic Space Sculpting

Byeongjoo Ahn, Michael De Zeeuw, Ioannis Gkioulekas, Aswin C. Sankaranarayanan

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023

project page · paper · video · code · bibtex

Full-surround 3D reconstructions from a single kaleidoscopic image via neural surface sculpting.
Kaleidoscopic Structured Light

Byeongjoo Ahn, Ioannis Gkioulekas, Aswin C. Sankaranarayanan

ACM Transactions on Graphics (Proc. SIGGRAPH Asia), 2021

project page · paper · video · code · bibtex

Structured light with hundreds of virtual projectors and cameras for full-surround scanning.
Convolutional Approximations to the General Non-Line-of-Sight Imaging Operator

Byeongjoo Ahn, Akshat Dave, Ashok Veeraraghavan, Ioannis Gkioulekas, Aswin C. Sankaranarayanan

IEEE/CVF International Conference on Computer Vision (ICCV), 2019 (Oral)

project page · paper · video · code · bibtex

A convolutional perspective of the NLOS operator enabling efficient reconstruction.
Occlusion-Aware Video Deblurring with a New Layered Blur Model

Byeongjoo Ahn, Tae Hyun Kim, Wonsik Kim, Kyoung Mu Lee

Tech report, 2016

paper · bibtex

Layered blur model handling occlusions for video deblurring.
Reduced Illumination Patterns for Acquisition of Specular and Diffuse Normal Maps

Byeongjoo Ahn, Junghyun Cho, Taekyung Yoo, Ig-Jae Kim

ACM SIGGRAPH Asia Posters, 2016

paper · bibtex

Polarization-based normal map acquisition with fewer patterns.
Dynamic Scene Deblurring

Tae Hyun Kim, Byeongjoo Ahn, Kyoung Mu Lee

IEEE International Conference on Computer Vision (ICCV), 2013

paper · bibtex

Joint estimation of latent image, motion, and soft segmentation for dynamic scenes.

Byeongjoo Ahn (안병주), Research Scientist at Apple

Professional Activities

Research