Guess That Animal Doodle

Description

This project is an AI application that recognizes drawings based on the Quick, Draw! dataset. The application uses a convolutional neural network (CNN) to predict the category of a doodle drawn on a 28 x 28 grid.

Installation Instructions (MacOS)

Prerequisites

• Python 3.x
• Pip (Python package installer)
• A virtual environment (recommended)

1. Clone the Repository

   git clone https://github.com/vukvukovich/guess-that-animal-doodle.git
   cd guess-that-animal-doodle

2. Create a Virtual Environment and install tensorflow

   Follow Apple guide from here.

3. Install the Required Packages

  pip install -r requirements.txt

4. Prepare Datasets

   Download the .npy files for different animal categories from the Google Quick, Draw! Dataset. Place them in the ml/datasets directory. Make sure to rename the files and naming convention is respected e.g. new-animal.npy.

Training the Model

1. Train the Model

  python ml/train.py

Running the Application

1. Start the Flask Application

  python app.py

2. Open Your Browser

Navigate to http://127.0.0.1:5000/ to access the application.

Project Structure

guess-that-animal-doodle/
├── app.py
├── ml
│   ├── datasets
│   │   ├── bear.npy
│   │   ├── bird.npy
│   │   ├── cat.npy
│   │   ├── cow.npy
│   │   ├── dog.npy
│   │   ├── elephant.npy
│   │   ├── fish.npy
│   │   ├── giraffe.npy
│   │   ├── horse.npy
│   │   ├── lion.npy
│   │   ├── monkey.npy
│   │   ├── mouse.npy
│   │   ├── pig.npy
│   │   └── sheep.npy
│   ├── models
│   │   └── quickdraw_cnn_model_some_animals.keras
│   └── train.py
├── static
│   ├── css
│   │   └── style.css
│   ├── images
│   │   └── logo.webp
│   └── js
│       └── script.js
├── templates
│   └── index.html
├── requirements.txt
└── README.md