|
Hyelin Nam I'm an AI Research Engineer at EverEx. Before joining EverEx, I completed my master's at KAIST, where I was advised by Jong Chul Ye. Recently, I have worked on diffusion models, leveraging their adaptability to enhance the quality of generated outputs. My goal is to advance generative models to accurately capture real-world dynamics, ultimately enabling them to function as reliable real-world priors. **I am actively looking for PhD openings starting in Fall 2025.** |
|
Research |
|
[C4]
Optical-Flow Guided Prompt Optimization for Coherent Video Generation
Hyelin Nam*, Jaemin Kim*, Dohun Lee, Jong Chul Ye CVPR, 2025 project page / arXiv / code Prompt optimization driven by an optical flow discriminator to enhance temporal consistency and natural motion dynamics in video diffusion models. |
|
[C3]
CFG++: Manifold-constrained Classifier Free Guidance For Diffusion Models
Hyungjin Chung*, Jeongsol Kim*, Geon Yeong Park*, Hyelin Nam*, Jong Chul Ye ICLR, 2025 Silver prize, 31st Samsung Humantech Paper Award project page / arXiv / code A simple fix to CFG that enables lower guidance scales, improves sample quality and invertibility. |
|
[C2]
Contrastive Denoising Score for Text-guided Latent Diffusion Image Editing
Hyelin Nam, Gihyun Kwon, Geon Yeong Park, Jong Chul Ye CVPR, 2024 project page / arXiv / code Ensure structural correspondence by leveraging diffusion features during the score distillation process. |
|
[C1]
HairFIT: Pose-invariant Hairstyle Transfer via Flow-based Hair Alignment and Semantic-region-aware Inpainting
Chaeyeon Chung*, Taewoo Kim*, Hyelin Nam*, Seunghwan Choi, Gyojung Gu, Sunghyun Park, Jaegul Choo BMVC, Oral Presentation, 2022 Best Paper Award, Korean Artificial Intelligence Association, 2021 arXiv |