Kanghyun Baek — accepted AI-conference papers

2026

Diagnosing and Correcting Concept Omission in Multimodal Diffusion Transformers

Kanghyun Baek, Jaihyun Lew, Chaehun Shin, Jungbeom Lee, Sungroh Yoon

ICML 2026poster

Multimodal Diffusion Transformers (MM-DiTs) have achieved remarkable progress in text-to-image generation, yet they frequently suffer from concept omission, where specified objects or attributes fail to emerge in the generated image. By performing linear probing on text tokens, we demonstrate that t…

Cited by 0 Source ↗Scholar ↗

2025

DefectFill: Realistic Defect Generation with Inpainting Diffusion Model for Visual Inspection

Jaewoo Song, Daemin Park, Kanghyun Baek, Sangyub Lee, Jooyoung Choi, Eunji Kim +1

CVPR 2025highlight

Developing effective visual inspection models remains challenging due to the scarcity of defect data. While image generation models have been used to synthesize defect images, producing highly realistic defects remains difficult. We propose DefectFill, a novel method for realistic defect generation…

Cited by 0 Source ↗PDF ↗Scholar ↗