stanislawbilewski / imagelabelingpipeline

ImageLabelingPipeline

Overview

This repository contains the ImageLabelingPipeline project, developed for an MLOps class at our university. The project consists of an application designed to help in labeling images in a dataset for machine learning.

Object Detection

Due to resource constraints, the project's focus is the manual labeling of data and creation of utilities to support machine learning operations rather than training a complex model.

Installation

Prerequisites:

• Python 3.x
• Flask
• cleanvision package

Clone the repository and navigate to the project directory. Install the required packages using the requirements.txt file provided:

pip install -r requirements.txt

Usage

1. Start the Flask app:

python app.py

2. Access the application through your web browser at localhost:5001.
3. Use the web interface to upload images for annotation.
4. Run clean_analysis.py to analyze the uploads for any unsuitable images using CleanVision.

Tools Utilized

• CleanVision: for dataset analysis and image selection.
• DVC: for data version control (not integrated in the given code).

Project Structure

• app.py: The Flask application's main entry point.
• static/: Contains CSS and JS files for the front-end.
• templates/: HTML templates for rendering the web interface.
• clean_analysis.py: A script to analyze and report issues in the uploaded images.

Contribution

Contributions to the project are welcome. Please ensure proper testing and documentation for any submitted features or bug fixes.

License

The project is open-sourced under the MIT license.