This repository contains a Convolutional Neural Network (CNN) based doodle classifier, trained to recognize 10 different doodle classes: star, shoe, piano, house, popsicle, tshirt, umbrella, duck, airplane, and windmill. The model is deployed as a web application using Flask, allowing users to upload their doodles and receive predictions.
doodle-classifier/
├── app.py # Flask application
├── keras.h5 # Saved model weights and architecture
├── templates/ # HTML templates for the web application
│ └── index.html
├── static/ # Static files (CSS, JavaScript, images)
│ └── style.css
│ └── sketch.js
│ └── jquery-1.3.2.min.js
└── README.md
The CNN model was trained using a dataset of doodles. The trained model weights are saved in models/doodle_classifier.h5.
To run the Flask application, execute the following command:
run flask
The application will start running on http://127.0.0.1:5000/. Open this URL in your web browser.
Open the web application in your browser.
You will see an interface to upload a doodle image.
Draw image of a doodle (star, shoe, piano, house, popsicle, tshirt, umbrella, duck, airplane, or windmill).
Click the "Predict" button.
The application will display the predicted class of the doodle.
Model Architecture: Convolutional Neural Network (CNN)
Classes: star, shoe, piano, house, popsicle, tshirt, umbrella, duck, airplane, windmill
Frameworks: Tensorflow, Keras, Flask
Improve model accuracy by using a larger and more diverse dataset.
Implement real-time doodle recognition.
Add more classes to the classifier.
Implement better error handling and user feedback.
Deploy the application on a cloud platform for wider accessibility.
Contributions are welcome! If you find any issues or have suggestions for improvements, please feel free to open an issue or submit a pull request.
Feel free to use the application and don't forget to give credits