pynicolas fe91f3e241 Improve distinction between color and grayscale documents (#79)
Better differentiate color and grayscale documents:
- Look for colored pixels only in the intersection of the mask and quadrilateral
- Apply a white balance (grey world) to the document
- Exclude pixels with extreme luminance
- Erode segmentation mask
2025-12-10 17:08:21 +01:00
2025-12-07 21:39:49 +01:00
2025-12-01 16:33:05 +01:00
2025-05-09 12:10:11 +02:00
2025-03-30 15:07:54 +02:00
2025-11-17 06:21:39 +01:00

FairScan icon

FairScan

An Android app to scan your documents —
simple and respectful.

License

Get it on: F-Droid · Google Play · GitHub


FairScan is an Android app to scan documents quickly, easily and privately.

It's designed to be simple: users get a clean, shareable PDF in seconds, with no manual adjustments.
And respectful: open source, minimal permissions, no tracking, no ads.


Screenshots

Scan Preview Save & Share

Features

  • Clear, distraction-free interface
  • Easy scanning process: scan → preview → share
  • Automatic document detection using a custom segmentation model
  • Automatic perspective correction
  • Automatic image enhancement
  • Fast PDF generation with no manual adjustments
  • Fully offline the app has no internet permission
  • Minimal permissions
  • Open source, GPLv3

Compatibility

FairScan works on any device that:

  • runs Android 8.0+
  • has a camera

Technical details

FairScan uses:


The segmentation model

FairScan uses a custom-trained image segmentation model to detect documents:
https://github.com/pynicolas/fairscan-segmentation-model

It's based on a fully public dataset that is available here:
https://github.com/pynicolas/fairscan-dataset

The build system automatically downloads the model using
download-tflite.gradle.kts.

Related blog posts:


Build

To build an APK:

./gradlew clean check assembleRelease

To build an Android App Bundle:

./gradlew clean check :app:bundleRelease

License

This project is licensed under the GNU GPLv3. See LICENSE for details.

Description
No description provided
Readme GPL-3.0 5 MiB
Languages
Kotlin 93.9%
Python 5.7%
xBase 0.4%