JunkWaxDetection.LiveDemo

JunkWaxDetection.LiveDemo is an ML-powered, real-time Blazor web application built on .NET 9 that detects Junk Wax Baseball Cards — including retro baseball cards from the 80s and 90s — using advanced machine learning models. This project leverages ML.NET, ONNX, and Paddle OCR to deliver precise image classification and optical character recognition (OCR) for authenticating and analyzing vintage baseball cards. My overall goal for this tech demo was to show that C# developers can have just as much fun with Machine Learning and Computer Vision as any other language!

Table of Contents

• Overview
• Key Features
• Technical Stack
• Prerequisites
• Getting Started
• Usage Instructions
• Machine Learning Integration
• Trading Card Datasets
• Contributing
• License

Overview

JunkWaxDetection.LiveDemo provides a robust solution for detecting junk wax baseball cards, with a special focus on retro baseball cards from the 80s and 90s. By integrating cutting-edge machine learning algorithms and computer vision techniques, this project helps collectors, dealers, and enthusiasts verify the authenticity and quality of their vintage baseball card collections.

Key Features

• Real-Time Junk Wax Detection: Instant identification of junk wax on baseball cards.
• Optimized for Retro Baseball Cards: Special considerations for vintage baseball cards from the 80s and 90s.
• Interactive UI: Modern user interface built with Blazor.
• Machine Learning Integration: Utilizes ML.NET and ONNX models for robust image classification.
• Optical Character Recognition (OCR): Uses Paddle OCR to extract text from baseball card images for deeper analysis.
• Data-Driven Insights: Integrates with open sports card datasets to enhance card authentication and verification processes.

Technical Stack

• .NET 9 SDK: Core framework powering the application.
• Blazor: Front-end framework for building interactive web UIs with C#.
• ML.NET: Machine learning framework for loading and executing the ONNX model.
• ONNX: Open Neural Network Exchange format for deploying pre-trained ML models.
• Paddle OCR: Advanced OCR tool for extracting text from images.
• Visual Studio 2022: Recommended IDE for development and debugging.

Prerequisites

• .NET 9 SDK
• Visual Studio 2022 or later

Getting Started

1. Clone the Repository

   git clone https://github.com/JunkWaxData/JunkWaxDetection.LiveDemo.git

2. Navigate to the Project Directory

   cd JunkWaxDetection.LiveDemo

3. Open the Solution
   Open JunkWaxDetection.LiveDemo.sln in Visual Studio 2022.
4. Restore Dependencies
   Run the following command in your terminal:

   dotnet restore

5. Run the Application
   Start debugging or run:

   dotnet run

Usage Instructions

• Open your browser and navigate to https://localhost:5001.
• Utilizing your webcam show your favorite retro baseball card from the 80s or 90s.
• The system processes the image using machine learning models to detect the card being presented.
• View detailed results, including OCR analysis and validation data.

Machine Learning Integration

Junk Wax Detection ONNX Model

I use the JunkWaxDetection ONNX Model for precise image classification. This model is integrated with ML.NET using the MLContext and PredictionEngine classes, ensuring efficient predictions and optimal performance for detecting junk wax baseball cards.

ML.NET and ONNX

• ML.NET: Simplifies the integration and execution of machine learning models within .NET applications.
• ONNX: Ensures scalability and cross-platform compatibility for deploying pre-trained models across various environments.

Paddle OCR for Optical Character Recognition

In addition to visual classification, Paddle OCR extracts textual data from baseball card images. This text extraction is essential for matching card details against vintage baseball card databases and further identifying retro cards from the 80s and 90s.

Trading Card Datasets

Junk Wax Data Card Lists

The project leverages the open-source Junk Wax Data Card Lists repository, which provides a comprehensive JSON dataset of sports trading cards. This dataset supports:

• Matching player names and card details extracted via OCR.
• Cross-referencing card information to enhance the validation process for vintage collections.

Contributing

Contributions are welcome! To contribute:

• Fork the repository.
• Create your feature branch:

  git checkout -b feature/your-feature-name

• Commit your changes:

  git commit -m 'Add some feature'

• Push to the branch:

  git push origin feature/your-feature-name

• Open a Pull Request for review.

Please refer to the repository guidelines for detailed contribution instructions.

License

This project is licensed under the MIT License. You are free to use, modify, and distribute this project under the terms of the license.