ROPSCI, Rock Paper Scissors 🪨📄✂️

This repository contains an interactive Rock Paper Scissors game implemented using computer vision techniques. The project leverages deep learning to recognize hand gestures in real-time, providing an engaging and interactive experience. This is the final project of the Computer Vision course.

📁 Project Structure

The project is organized as follows:

├── backend/             # Backend of the project written with Django
├── frontend/            # Frontend of the project written with Nextjs
├── images/              # Readme images
├── README.md            # Main README file
└── LICENSE              # License

🎯 Project Overview

The goal of this project is to implement an interactive Rock-Paper-Scissors game using computer vision. The system detects hand gestures via a deep learning model (YOLOv11) and displays the game results in real time.

Key Features:

• 🎥 Real-time hand gesture recognition
• 🎮 Intuitive user interface
• 🚫 Anti-cheating mechanisms
• 👥 Multiplayer mode with WebSockets
• ⚡ Optimized performance for smooth gameplay

🛠️ Tech Stack

Frontend (User Interface)

• Next.js and React for a dynamic and responsive UI
• WebSocket integration for real-time multiplayer communication

Backend (Game Logic & API)

• Django & Django Channels for managing game logic
• WebSocket support for real-time communication
• YOLOv11 for hand gesture recognition
• Redis for message brokering and fast state management

Machine Learning Model

• YOLOv11 trained on a dataset of hand gestures (rock, paper, scissors)
• Data augmentation for improved accuracy
• Real-time inference for smooth gameplay

🚀 Installation & Setup

First, copy one of your desired models from the models folder to the backend folder and rename it to model.pt. Then, execute these commands.

git clone https://github.com/mfrashidi/ropsci.git
cd ropsci
cd backend
python -m venv venv
source venv/bin/activate   # On Windows, use `venv\Scripts\activate`
pip install -r requirements.txt
python manage.py makemigrations
python manage.py migrate
daphne ropsci.asgi:application -b localhost -p 8000 # Websocket
daphne ropsci.asgi:application -b localhost -p 8001 # API
cd ../frontend
npm i
npm run dev

Now you can play the game from http://localhost:3000

---

🧪 Dataset & Model Training

Dataset Collection

• Over 1000 images were collected for each hand gesture (rock, paper, scissors) from different angles and lighting conditions.
• Data augmentation techniques (rotation, brightness adjustment) were applied to enhance model performance.

Model Training

• Used YOLOv11 for real-time object detection.
• Roboflow was used to preprocess and annotate the dataset.
• The model was trained with 150 epochs, achieving high accuracy.
• Small and nano versions of YOLO were tested for performance vs accuracy.

🎮 Game Features

Multiplayer Support

• Players can create a room and invite friends using a unique game code.
• The game is synchronized in real-time via WebSockets.

Anti-Cheat System

• The game detects fake movements and prevents last-second changes.
• A red mask is applied to cheaters to indicate a rule violation.

Winner Announcement

• The final winner is announced at the end of the rounds.
• A golden crown is displayed on the winner's avatar.

3D Models (Crown & Mask)

• Three.js and Mediapipe were used to overlay 3D objects (a crown for winners and a mask for cheaters).
• Lightweight OBJ models were loaded dynamically for optimized rendering.

---

This project was developed by @mfrashidi and @Bahram03 and @Arshiapdr. We hope you enjoy playing Rock-Paper-Scissors with our interactive computer vision game! 🎉