/Chess-Game-State-Detection

A computer vision system that detects state of a chess game by performing chess board detection and piece classification.

Primary LanguagePython

ChessGameStateDetectionCV

Deadline: 17.12.2022

Submission Requirements:

Code

Report

  • Used algorithms
  • Experiment results and analysis
  • Performance & accuracy (refer to experiment results section below)
  • Conclusion and references.
  • Work division between team members.
  • Any additional comments

Experiment results

  • Variety of testcases
  • Choice of comparison metric: accuracy, recall and/or etc.
  • Strengths and weaknesses

Testcases

Readme file

  • Thats me, hello! 👋

How to run the project

Requirements

TODO

How to run

Change the config file to your needs and add the images you need to test in the dir folder chosen in the config file and run the main.py file.

Configuration

TODO

Code structure

TODO

TODO:

Detection:

  • Grayscale
  • Canny
  • Thershold
  • Hough
  • H-space
  • Outlier elimination
  • Create reference img
  • Find corresponding intersections
  • Define reference img
  • Predict intersections
  • Geometeric projection

Classification

  • Create dataset
  • HOG feature extraction
  • Train model
  • Test model
  • Evaluate model

Bonus

  • Fancy GUI
  • Converting idea into real-time application
  • Unique features
  • High Speed-up performance

References