2048 Game Assistant

Developed an Android app that analyzes live 2048 game state and suggests optimal moves

Overview

Created an intelligent assistant for the game 2048 that combines real-time image processing with AI decision-making to analyze the game state and recommend optimal moves.

GitHub Repository
2048 Game Assistant

System Architecture

System architecture diagram showing the image processing pipeline and AI decision-making process.

Technical Details

  • Implemented advanced image processing pipeline:
    • Grayscale conversion
    • Canny edge detection
    • Perspective transformation
    • Template matching for digit recognition
  • Developed AI decision engine:
    • Expecti-max Search algorithm
    • Corner placement strategy
    • Move evaluation heuristics
  • Built Android application:
    • Real-time camera feed processing
    • Game state analysis
    • Move recommendation display

Key Achievements

  • 100% digit recognition accuracy with proper alignment
  • Consistent AI performance reaching 1024 tile in 75% of simulations
  • Real-time processing and move suggestion
  • Robust board state detection
  • Efficient template matching system

Technical Stack

  • Android Studio
  • Java/Kotlin
  • OpenCV
  • Image Processing
  • Machine Learning
  • Real-time Systems
  • Mobile Development
  • Computer Vision

Project Advisor

Prof. Thomas Moon, UIUC

Demo of our ECE420 Final Porject -- "Game 2048 Cheat Machine" (Right click on the cover above to view the video)