Publications

You can also find my articles on my Google Scholar profile.

Conference (full paper)

On Epistemic Uncertainty of Visual Tokens for Object Hallucinations in Large Vision-Language Models

Published in Neural Information Processing Systems (NeurIPS), 2025

This paper is about mitigating hallucinations of Large Cision-Language Models via training-free method.

Authors: Hoigi Seo*, Dong Un Kang*, Hyunjin Cho, Joohoon Lee, and Se Young Chun. (* co-first author)
Download Paper

Skrr: Skip and Re-use Text Encoder Layers for Memory Efficient Text-to-Image Generation

Published in International Conference on Machine Learning (ICML), 2025

This paper is about pruning large text encoder in T2I diffusion model for memory efficient image synthesis.

Authors: Hoigi Seo*, Wongi Jeong*, Jae-sun Seo, and Se Young Chun. (* co-first author)
Download Paper

Efficient Personalization of Quantized Diffusion Model without Backpropagation

Published in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), The 2nd Workshop on Efficient and On-Device Generation (EDGE), 2025

This paper is about personalizing Texto-to-Image diffusion models with extreme memory efficiency via zero-th order optimization of quantized model.

Authors: Hoigi Seo*, Wongi Jeong*, Kyungryeol Lee, and Se Young Chun. (* co-first author)
Download Paper

INTRA: Interaction Relationship-aware Weakly Supervised Affordance Grounding

Published in European Conference on Computer Vision (ECCV), Twelfth International Workshop on Assistive Computer Vision and Robotics (ACVR24) (oral), 2024

This paper is about training models to tackle affordance grounding with weakly supervised learning.

Authors: Ji Ha Jang*, Hoigi Seo*, and Se Young Chun. (* co-first author)
Download Paper

BeyondScene: Higher-Resolution Human-Centric Scene Generation With Pretrained Diffusion

Published in European Conference on Computer Vision (ECCV), 2024

This paper is about generating human-centric ultra-high resolution image with pre-trained diffusion model.

Authors: Gwanghyun Kim*, Hayeon Kim*, Hoigi Seo*, Dong Un Kang*, and Se Young Chun. (* co-first author)
Download Paper

Conference (short paper, abstracts or workshop papers)

Harmonization for a black-box deep learning model

Published in International Society for Magnetic Resonance in Medicine (ISMRM) (oral, Summa cum laude), 2025

This paper is about harmonizing MR image for a black-box deep learning model

Authors: Minjun Kim, Hwihun Jeong, Hoigi Seo, Se Young Chun, and Jongho Lee

Technical Report

Upsample What Matters: Region-Adaptive Latent Sampling for Accelerated Diffusion Transformers

Published in arXiv, 2025

This paper is about accelerating Text-to-Image diffusion models with region-adaptive upsampling

Authors: Wongi Jeong*, Kyeongryeol Lee*, Hoigi Seo, and Se Young Chun. (* co-first author)
Download Paper

On Geometrical Properties of Text Token Embeddings for Strong Semantic Binding in Text-to-Image Generation

Published in arXiv, 2025

This paper is about analyzing its geometrical properties in cross-attention and leverage its properties for strong semantic binding.

Authors: Hoigi Seo*, Junseo Bang*, Haechang Lee*, Joohoon Lee, Byung Hyun Lee, and Se Young Chun. (* co-first author)
Download Paper

DITTO-NeRF: Diffusion-based Iterative Text To Omni-directional 3D Model

Published in arXiv, 2023

This paper is about generate diverse 3D object with pre-trained 2D diffusion model.

Authors: Hoigi Seo*, Hayeon Kim, Gwanghyun Kim and Se Young Chun. (* co-first author)
Download Paper