picture

Daeyun Shin


My research interests are in artificial intelligence at the intersection of computer vision and computer graphics. I am interested in 3D scene understanding and unsupervised learning. I am advised by Charless Fowlkes.

Education

Since 2017, Ph.D. in Computer Science, University of California, Irvine (advisor: Charless Fowlkes)

2017, M.S. in Computer Science, University of Illinois at Urbana-Champaign (advisor: Derek Hoiem)

2015, B.S. in Computer Science, University of Illinois at Urbana-Champaign

Publications

Pixels, voxels, and views: A study of shape representations for single view 3D object shape prediction
Daeyun Shin, Charless Fowlkes, Derek Hoiem
In CVPR, 2018   [pdf]   [supplementary material]   [poster]   [code]   [project website]

3DFS: Deformable Dense Depth Fusion and Segmentation for Object Reconstruction from a Handheld Camera
Tanmay Gupta, Daeyun Shin, Naren Sivagnanadasan, Derek Hoiem
In arXiv:1606.05002, 2016   [arxiv]   [video results]

Completing 3D Object Shape from One Depth Image
Jason Rock, Tanmay Gupta, Justin Thorsen, Junyoung Gwak, Daeyun Shin, Derek Hoiem
In CVPR, 2015   [pdf]   [project website]

Current topics

  • Multi-view 3D volume learning and reconstruction.
  • Learning novel objects.
  • Multi-task learning.
  • Geometry and scene understanding.

Projects, Code, tools

  • Python library for vision research Code

  • Multi-view 3D shape graph cut implementation (Campbell et al.) Code

  • Improving Randomzied SVD Code Report

  • Off-screen depth and silhouette rendering using OSMesa. Python C

  • Saliency for 3D meshes Code Demo

  • Extracting informative graphics from packaging designs code/demo

  • Modern OpenGL renderer C++ GLSL shader

  • Image annotation toolkit code

  • MATLAB utils
    • Load Google Earth KMZ (.kmz) file as a triangular mesh. code
    • Exporting figures code
  • Vim and Unix configurations .vimrc, others

  • LaTeX templates

  • Other stuff on Github: link