Doodle Recognition with MobileNet and OpenCV

This project implements doodle recognition using a MobileNet model trained on the Google QuickDraw dataset. It allows users to perform recognition on imported images.

Data found here

Features

• Train a MobileNet model for doodle recognition using the QuickDraw dataset.
• Recognize human-drawn doodles from .png and .jpg images

Dependencies

• Python 3.x
• TensorFlow
• NumPy
• OpenCV
• Pillow (optional, used for image processing)
• Matplotlib (optional, used for data visualization)

Usage

1. Clone the repository: git clone https://github.com/AadiJo/Doodle-Recognition.git

2. Navigate to the project directory: cd 'your-repo'

3. Install the dependencies: pip install -r requirements.txt

4. Run the main script: python gui.py

5. Import image: Click the icon in the top left to select an image. File format is 32 x 32, white drawing on black background

6. Detect image: Click the detect button!