Virginia Tech®home

Object detection on ETD (Electronic Theses and Dissertations)

Object detection on ETD (Electronic Theses and Dissertations)
Loading player for https://youtu.be/4kfcOyXceKw...

We would like to extract different elements of scholarly documents such as metadata (title, authors, year), chapter headings and subheadings, equations, figures (and captions), tables (and captions), and paragraphs.

We build a classifier to detect bounding boxes for each of these types of objects. We also annotate the dataset with bounding boxes for each of the elements, and use detectron2 for training the object detection model.

Organization

  • CS4624 Object Detection team

Team

  • Kecheng Zhu
  • Jiangyue li
  • Zachary Gager
  • You peng
  • Shelby Neal