Skip to content

Convolutional Neural Network trained to classify hand-drawn images

Notifications You must be signed in to change notification settings

Shynee1/ImageGuesser

Repository files navigation

Image Guesser

Want to test your drawing skills? See if my neural network can guess what you are drawing!
This model has been trained on Google's QuickDraw dataset to recognize hand-drawn doodles

Demonstration

Table of Contents

Features

The CNN can recognize 15 different types of drawings:

  • Banana
  • Bandage
  • Book
  • Cake
  • Camera
  • Cell Phone
  • Face
  • Hamburger
  • Pizza
  • Sailboat
  • Sheep
  • Star
  • Sword
  • Television
  • Effiel Tower

Setup & Installation

  1. Clone the repository:
    $ git clone https://github.com/Shynee1/ImageGuesser.git
    $ cd ImageGuesser
  2. Install dependencies:
    $ pip install -r requirements.txt
  3. Load dataset:
    $ python load_data.py
  4. Train model:
    $ python model.py
  5. Run program
    $ python draw.py

Project Information

This project was made as my final project for Brown University's summer AI program.

About

Convolutional Neural Network trained to classify hand-drawn images

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages