Skip to main content

Layout Segmentation Simplifier

Project description

LSS (Layout Segmentation Simplifier)

Install

pip install https://github.com/PonteIneptique/lss/archive/refs/heads/main.zip

or

pip install layouts-segmentation-simplifier

Use

Generic use

from lss.parsers import PageXML

file = PageXML.from_file(
    # Path to your file
    "data/0002_Main_frame.xml",
    # Optional: set-up the namespace, as they tend to change a lot
    namespace="http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15" 
)
# Simplify your baseline: things within 10% of your line height will be discarded (Seems to be a good number)
file.simplify_lines(ratio=.10)
# Simplify your baseline: things within 15% of your mask height will be discarded (Seems to be a good number)
file.simplify_masks(ratio=.15)
# You can retrieve the xml in file.xml or dump the modified xml in file.xml
file.dump(filepath="file.xml")

Qualitatively search for best parameters

You can also get it to make multiple simulation, to chose the best value (highest ratio = bigger simplification)

from lss.parsers import PageXML

page = PageXML.from_file("./data/0002_Main_frame.xml")
page.find_namespace() # Useful in case you use PageXML as it scans for the right namespace
page.test_values(
    [(.05, .05), (.1, .1), (.15, .15), (.20, .20)],
    image="./data/0002_Main_frame.jpg",
    basename_output="./data/0002_Main_frame"
)

In this context, lines will be simplified by a ratio of 0.05, 0.10, 0.15 and 0.20. Mask will be as well. Few files will be saved. Each test will create test files with overlays (cf. folder data for all examples):

Original 15% ratio on both line and masks
Original Mask Original Mask
10% 20%
Original Mask Original Mask

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

layouts-segmentation-simplifier-0.0.2.tar.gz (9.0 kB view hashes)

Uploaded Source

Built Distribution

layouts_segmentation_simplifier-0.0.2-py2.py3-none-any.whl (8.8 kB view hashes)

Uploaded Python 2 Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page