No description, website, or topics provided.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
detection
experiments
ext/notebook
image_sampling
mask-rcnn
.gitignore
README.md
__init__.py
meeting_agendas.md
preproc.py
sampler.py

README.md

Book Annotation Classification Engine

A UCLA Collections Lab/BuildUCLA project: computer vision experiments to identify and classify annotations in digitized printed books in IIIF-hosted collections.

Status: Experimenting with different neural network architectures and evaluating their performances on a dataset of over 3000 images.

Want to help us build our training data? Tag and classify annotation here: https://www.zooniverse.org/projects/kirschbombe/book-annotation-classification

Data

https://ucla.app.box.com/folder/45481483089

Project team

  • Dawn Childress
  • Pete Broadwell
  • Andrew Wallace
  • Johnny Ho
  • Emily Chen
  • Jonathan Quach
  • Morgan Madjukie
  • Rahul Malavalli

More Info on Some Files in this Path

.gitignore

Intentionally not track pycache .pyc and .ipynb_checkpoints since they do not provide convenient information about changes that may have occured to the code.

init.py

Standard file to be included to treat folders as packages (list of modules) so as to allow modularization

preproc.py

Binarizes an image or a folder of images (NOTE: must provide an output path if binarizing an entire folder)

E.g.

$ python preproc.py img.png #one image $ python preproc.py --out /Users/John/Desktop/outputFolder /Users/John/Desktop/inputFolder

sampler.py

Randomly generates "number" square subsamples of side length "px" from file(s) "files" where "number" "px" and "files" are from user input.