QR Code Detection, Cropping, and Decoding using Machine Learning

This repository contains code and resources for a machine learning model that detects, crops, and decodes QR codes from images using a Jupyter Notebook.

Overview

QR (Quick Response) codes are two-dimensional barcodes that store various types of information. This project focuses on creating a machine learning model to automatically detect QR codes within images, crop them for better isolation, and decode the information contained in the QR codes.

Repository Structure

The repository is organized as follows:

• datasets/: Main dataset used to train the model, with images separated into test, train, and validate folders.
• datasets 0/: Previous dataset used to train the model, with images separated into test, train, and validate folders.
• Images/: Pre-processed dataset of images with and without QR codes.
• models/: Previous models.
• QR-codes/: Folder to save cropped QR codes from images.
• runs/: Folder containing all previous machine learning models, with the final one located in train 12/.
• QR-detector.ipynb: Jupyter Notebook containing the code for the project.
• IMG_20230820_100924_960.jpg: Image used for manual testing of the model.

Getting Started

1. Clone the repository to your local machine:

git clone https:github.comstefano-lacorazza/QR_Code_Detector_Notebook.git
cd QR_Code_Detector_Notebook.

2. Upload image you wish to evaluate.
3. Open QR_detector.ipynb.
4. Change source in cell 9 to your image name.
5. Run cell 37, to install packages locally.
6. Run cell 9, it will show the detected QR codes.
7. Run cell 49, the cropped QR codes will be in the QR-codes directory.
8. Run cell 21 to display the decoded url.
9. Run cell 22 to see a preview of the website.