Dong Ju Mun.

I am a Ph.D. candidate working at the intersection of diffusion models and generative AI. I build generative methods, grounded in diffusion and flow models, for solving inverse problems in imaging, advised by Prof. Se Young Chun at the Intelligent Computational Imaging Lab, Seoul National University.

My research connects the sampling dynamics of generative models with the fidelity that real measurements demand. I am interested in turning expressive generative priors into reliable, faithful reconstructors, and in bringing them to imaging systems that span natural images, lensless cameras, and modern CMOS image sensors.

💡 Research interest: turning generative priors into reliable solvers for ill-posed inverse problems. My recent work studies how jointly scheduling guidance and stochasticity stabilizes diffusion posterior sampling, how training-free diffusion can faithfully restore measurements under severe, complex degradation such as lensless imaging, and how zero-shot diffusion extends to raw image restoration and demosaicing across Bayer and non-Bayer sensors.

Publications

Research

Publications

all publications →

2026

ICML 2026

Triadic Dynamics Aware Diffusion Posterior Sampling for Inverse Problems: Optimizing Guidance and Stochasticity Schedules

Junseo Bang*, Dong Ju Mun*, Hoigi Seo, Seongmin Hong, Se Young Chun

International Conference on Machine Learning (ICML), 2026

arXiv Code Project

@inproceedings{bang2026triadic,
  title  = {Triadic Dynamics Aware Diffusion Posterior Sampling for
            Inverse Problems: Optimizing Guidance and Stochasticity Schedules},
  author = {Bang, Junseo and Mun, Dong Ju and Seo, Hoigi and
            Hong, Seongmin and Chun, Se Young},
  booktitle = {International Conference on Machine Learning (ICML)},
  year   = {2026}
}

2025

IMAVIS 2025

Fidelity-Preserving Zero-Shot Diffusion Models for Highly Ill-Posed Inverse Problems in Lensless Imaging

Haechang Lee*, Dong Ju Mun*, Hyunwoo Lee*, Kyung Chul Lee, Seongmin Hong, Gwanghyun Kim, Seung Ah Lee, Se Young Chun

Image and Vision Computing (IMAVIS), 2025

Paper Code

@article{lee2025fidelity,
  title  = {Fidelity-Preserving Zero-Shot Diffusion Models for Highly
            Ill-Posed Inverse Problems in Lensless Imaging},
  author = {Lee, Haechang and Mun, Dong Ju and Lee, Hyunwoo and
            Lee, Kyung Chul and Hong, Seongmin and Kim, Gwanghyun and
            Lee, Seung Ah and Chun, Se Young},
  journal = {Image and Vision Computing},
  year   = {2025}
}

ICEIC 2025

Zero-Shot Diffusion Models for Demosaicing in Bayer and Non-Bayer Image Sensors

Haechang Lee, Dong Ju Mun, Joohoon Lee, Hyunwoo Lee, Se Young Chun

International Conference on Electronics, Information, and Communication (ICEIC), Jan. 2025

Paper

@inproceedings{lee2025zeroshot,
  title  = {Zero-Shot Diffusion Models for Demosaicing in
            Bayer and Non-Bayer Image Sensors},
  author = {Lee, Haechang and Mun, Dong Ju and Lee, Joohoon and
            Lee, Hyunwoo and Chun, Se Young},
  booktitle = {International Conference on Electronics, Information,
               and Communication (ICEIC)},
  year   = {2025}
}