Skip to content

Latest commit

 

History

History
100 lines (74 loc) · 2.98 KB

README.md

File metadata and controls

100 lines (74 loc) · 2.98 KB

Name

Card Master

Description

Card Master is a computer vision pipeline that helps identify the type and strength of playing cards using deep learning. By breaking down card features into separate classification tasks (color, suit, and rank), this project enables reliable card recognition even with limited training data. The final goal is to estimate the player’s winning probability using Monte Carlo simulation. It is built with Raspberry Pi deployment in mind for future wearable integration.

Badges

Badges (e.g., CI/CD status, code quality) to be added here.

Visuals

Some resources on Blackjack rules and strategies:

Installation

Datasets:

Requirements

  • Python 3.10+
  • TensorFlow
  • OpenCV
  • Treys
  • NumPy
  • scikit-learn
  • pickle

Instructions

Clone the repository and install dependencies:

pip install -r libraries.txt

Download the pre-trained models from the card_Master/thoughtful_approach folder for accurate predictions.

Usage

Predicting Card Attributes

predict_card_attributes(img_path)
  • Input: Path to a playing card image
  • Output: Dictionary with card color, suit, and rank
  • Internally calls separate CNN models for color, suit (depending on red/black), and rank detection

Estimating Win Rate

evaluate_win_rate(player_card_paths, num_simulations=1000, num_opponents=1)
  • Input: List of card image paths, number of simulations, number of opponents
  • Output: Estimated win rate as a float
  • Internally calls predict_card_attributes and simulates poker rounds using treys.Evaluator

Support

For help, please contact the author via GitLab Issues or direct message.

Roadmap

  • Integrate with Raspberry Pi camera module
  • Convert model inference pipeline to TensorFlow Lite
  • Build a wearable heads-up display for live card analysis
  • Expand model set to handle multi-card stacked images
  • Improve dataset with synthetic augmentations

Contributing

Contributions are welcome! Please fork the repository and open a merge request.
Include a brief description of your change and test it before submitting.

Authors and acknowledgment

Main Author: Zizhuo Yun
Built with support from ETH Zürich’s Python for Science & ML course team.

License

This project is for academic and educational purposes only. License terms to be added.

Project status

Ongoing development
The project is actively being improved and expanded. Contributions and feedback are welcome.