enzoribeirodev

enzoribeirodev / petrou

Public

A modular Python library for image thresholding and segmentation optimization.

16
0
89% credibility
Found May 21, 2026 at 16 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Petrou is a Python library that helps you automatically separate objects from backgrounds in grayscale images. It works by finding the optimal brightness cutoff point that divides your image into distinct regions. The library offers three different mathematical approaches (Otsu, Tsallis, and MASI) for finding this cutoff, and uses intelligent search strategies to locate the best threshold even in complex images. You can use it with default settings for quick results, or fine-tune the parameters for specialized images. The library also includes tools to measure how accurate your segmentation is against known correct answers, and can extract intensity profiles along lines in your images.

How It Works

1
🔬 You have images that need cleaning up

You have a collection of images where you need to separate the important parts from the background, like isolating cells in a microscope image or removing the background from a photo.

2
📦 You install the library

You add petrou to your project with a simple command, and it quietly handles all the heavy mathematical lifting behind the scenes.

3
🖼️ You load your grayscale image

You bring your image into the program as a simple grid of brightness values, just like opening a photo in any image viewer.

4
✨ The library finds the perfect cutoff point

Without you doing any math, the library examines your image's brightness distribution and discovers exactly where to draw the line between foreground and background.

5
You can let the library choose automatically, or fine-tune it yourself
🎯
Automatic mode

The library estimates the best parameters on its own, giving you a great result with zero extra work.

🔧
Manual mode

You provide specific numbers to guide the search, giving you more control over the final result.

6
📊 You check how well it worked

If you have a reference image showing the correct answer, the library measures your result against it and tells you the accuracy score.

🎉 Your image is perfectly segmented

You now have a clean binary image where the objects you care about are clearly separated from the background, ready for further analysis or use.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 16 to 16 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is petrou?

Petrou is a Python library for image thresholding and segmentation optimization. It gives you three well-established thresholding criteria (Otsu, Tsallis, and MASI entropy) with three different ways to find the optimal threshold: brute-force exhaustive search, simulated annealing, or particle swarm optimization. You also get a full suite of segmentation metrics like Jaccard index and Dice coefficient to evaluate how well your results match ground truth. The library uses a unified interface called SearchSpace that lets all optimizers work with any combination of float and integer variables without you having to write special-case code.

Why is it gaining traction?

The modular design means you can swap optimizers or add new thresholding criteria without rewriting boilerplate. The SearchSpace abstraction is particularly clever—it handles the messy details of perturbing mixed-type variables (floats vs integers) so your objective functions stay clean. The library also handles the tedious parts automatically: MASI and Tsallis parameters can be estimated from image statistics or jointly optimized with the threshold itself. If you need to extend it, the developer guide walks you through adding new optimizers or criteria in a few consistent steps.

Who should use this?

Computer vision engineers working on medical imaging, industrial inspection, or document analysis will find the most value here. Researchers benchmarking thresholding algorithms will appreciate having Otsu, Tsallis, and MASI under one roof with consistent APIs and built-in evaluation metrics. Anyone doing multi-level segmentation (producing more than just binary output) will benefit from the recursive splitting engine that works with any criterion.

Verdict

Petrou is a well-architected, genuinely modular Python library that solves a real niche problem cleanly. The credibility score sits at 0.8999999761581421%, reflecting its early stage (16 stars) and limited production history. The documentation is thorough and the code is well-structured, but test coverage and community size mean you should verify it works for your specific use case before betting on it in production. Worth trying for research or prototyping; hold off on mission-critical deployment until it builds more track record.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.