Hoigi Seo

Hoigi Seo

I am MS-PhD student majoring in Electrical and Computer Engineering at Seoul National University. My research interests are computer vision, gen AI, multi-model LLMs.

Hoigi Seo

Hi! :raised_hand: I am Hoigi Seo, a passionate researcher in the cutting-edge Artificial Intelligence (AI) field. I am also a MS-PhD student studying with marvelous colleagues in ICL, Seoul National University. My works focus on integrating multiple modal inputs with deep-learning models, diffusion-based generative model, Large Language Models (LLMs), and innate structure in vision problems. If you have any question about me or my publications, don’t hesitate to contact me!

News

  • [2025.04.01] My paper "Efficient Personalization of Quantized Diffusion Model without Backpropagation" is accepted to 2nd Workshop on Efficient and On-Device Generation (EDGE, CVPRW25)
  • [2025.02.27] My paper "Efficient Personalization of Quantized Diffusion Model without Backpropagation" is accepted to IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)! It will be presented as a poster. See you in Nashville!
    [Paper] [Project page] [Code]
  • [2025.02.12] My paper "Skrr: Skip and Re-use Text Encoder Layers for Memory Efficient Text-to-Image Generation" is opened to public!
    [Paper]
  • [2025.02.01] My paper "Harmonization for a black-box deep learning model" is accepted to International Society for Magnetic Resonance in Medicine (ISMRM25)! It will pe presented as an oral presentation.
  • [2024.09.20] My paper "INTRA: Interaction Relationship-aware Weakly Supervised Affordance Grounding" is accepted to Twelfth International Workshop on Assistive Computer Vision and Robotics (ECCVW24)! It will pe presented as an oral presentation in abstract track.
  • [2024.09.10] My paper "INTRA: Interaction Relationship-aware Weakly Supervised Affordance Grounding" is accepted to European Conference on Computer Vision 2024 (ECCV24)! It will pe presented as a poster.
    [Paper] [Project page] [Code]
  • [2024.09.10] My paper "BeyondScene: Higher-Resolution Human-Centric Scene Generation With Pretrained Diffusion" is accepted to European Conference on Computer Vision 2024 (ECCV24)! It will pe presented as a poster!
    [Paper] [Project page] [Code]
  • [2023.04.10] My paper DITTO-NeRF: Diffusion-based Iterative Text To Omni-directional 3D Model is published on arXiv!
    [Paper] [Project page] [Code]