Posts by Collection

portfolio

Portfolio item number 1

Short description of portfolio item number 1

Portfolio item number 2

Short description of portfolio item number 2

publications

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

Published in arXiv, 2023

Reconstructing diverse 3D objects from text or a single image with 2D diffusion priors.

Authors: Hoigi Seo*, Hayeon Kim, Gwanghyun Kim 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

Generating detailed, human-centric scenes beyond 8K with pretrained diffusion models.

Authors: Gwanghyun Kim*, Hayeon Kim*, Hoigi Seo*, Dong Un Kang* 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

Learning interaction-aware affordance grounding from weak supervision.

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

Harmonization for a black-box deep learning model

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

Harmonizing MR images for black-box models without access to model parameters.

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

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

Personalizing quantized diffusion models without backpropagation.

Authors: Hoigi Seo*, Wongi Jeong*, Kyungryeol Lee 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

Using text embedding geometry to improve semantic binding in text-to-image generation.

Authors: Hoigi Seo*, Junseo Bang*, Haechang Lee*, Joohoon Lee, Byung Hyun 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

Skipping and reusing text encoder layers for memory-efficient text-to-image generation.

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

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

Published in Neural Information Processing Systems (NeurIPS), 2025

Identifying and masking uncertain visual tokens to reduce object hallucinations without training.

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

Erasing Thousands of Concepts: Towards Scalable and Practical Concept Erasure for Text-to-Image Diffusion Model

Published in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (Compute Transparency Champion Award), 2026

Scalable concept erasure that removes thousands of concepts while preserving generation quality.

Authors: Hoigi Seo*, Byung Hyun Lee*, Jaehyun Cho, Sungjin Lim and Se Young Chun. (* co-first author)
Download Paper

Perceptually Consistent Low-Resolution Previews with High-Resolution Image for Efficient Workflows of Diffusion Models

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

Training-free, low-resolution previews that stay perceptually consistent with high-resolution outputs.

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

Training-free Mixed-Resolution Latent Upsampling for Spatially Accelerated Diffusion Transformers

Published in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (Highlight · Top 3.4%), 2026

Region-adaptive latent upsampling accelerates diffusion transformers with minimal quality degradation.

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

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

Published in International Conference on Machine Learning (ICML), 2026

Coordinating guidance and stochasticity schedules for diffusion-based inverse problems.

Authors: Junseo Bang*, Dongju Moon*, Hoigi Seo, Seongmin Hong and Se Young Chun. (* co-first author)
Download Paper

Strong Helps Weak: Directional Cross-Modal Alignment Transfer in Multi-modal LLMs

Published in Neural Information Processing Systems (NeurIPS) (Spotlight · Top 1.3%), 2026

Transferring alignment from strong vision models to audio and video MLLMs without further fine-tuning.

Authors: Hoigi Seo*, Byung Hyun Lee*, Minjun Kim*, Dohyun Mah, Jongho Lee and Se Young Chun. (* co-first author)
Download Paper

Recovering Diversity in Distilled Diffusion Models via Low-Rank Policy-Optimized Merging

Published in IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2026

Recovering diversity in few-step diffusion models through low-rank, policy-optimized teacher-student merging.

Authors: Hoigi Seo, Woojae Han, Joohoon Lee and Se Young Chun.

teaching

Teaching experience 1

Undergraduate course, University 1, Department, 2014

This is a description of a teaching experience. You can use markdown like any other post.

Teaching experience 2

Workshop, University 1, Department, 2015