Skip to content
A tool to mask text from screenshots using Azure Computer Vision API & ImageSharp
Branch: master
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.
src
.gitattributes
.gitignore
LICENSE.md
ProjectOxfordFun.sln
README.md
example-ndepend-crop.png
logo.svg

README.md

Trash PaNDA icon Trash PaNDA

Example of masked text in NDepend screenshot

What is Trash PaNDA?

Trash PaNDA is a tool to mask text from your images using SixLabors' ImageSharp and Microsoft's Computer Vision API.
Think of it as an automated redaction tool - just like you see with the odd Government report or statement. I initially designed it to be used to mask client names from screenshots of my code and tool windows, to comply with non-disclosure agreements (NDAs).

How do I use it?

Download with Chocolatey: choco install trashpanda Use in your own projects with Nuget: Install-Package TrashPaNDA Or check out the Releases tab for nightlies.

  1. Download, build the src\TrashPaNDA\TrashPaNDA.csproj.
  2. Call TrashPandaHelper.ReplaceWordsInImage():
    • string apiKey - your Cognitive Services API key;
    • byte[] originalImage - your original image stream in a byte array (ImageSharp will handle PNG, JPEG, GIF and BMP);
    • string[] wordsToReplace - an array of words to replace (case-insensitive);
    • int scaleUp - an optional scale factor to improve text recognition (leave as 1 & it'll scale automatically);
    • 'string endpoint' - optional, used to specify a specific Computer Vision API endpoint URI (see here for a list of acceptable endpoints)
  3. Do what you want with the resulting byte[] - MemoryStream, save to file, etc.

Contributing

PR's always accepted!

Credits

ImageSharp is licensed under an Apache 2.0 License.
Microsoft Cognitive Services API is licensed under an MIT License.
The Trash PaNDA logo is courtesy of Danielle Papanikolaou under the Pixabay License.

You can’t perform that action at this time.