HTML, CSS (SCSS/Sass) and JavaScript (jQuery) interactve and responsive front-end UI on top of a Python (Flask, Flask-RESTful, Flask-Caching, Flask-FlatPages, Flask-SQLAlchemy) back-end API and database*, implementing Python (numpy, sklearn, opencv) machine learning model building** and image recognition preprocessing*** via asyncronous encoded requests. Integrated with Git source control.

• * Database automatically clears itself up only when the application is booted, therefore it is not adviced to delete stored files whilist the processes are still running.

• ** Predictions are carried out by analyzing individual pixels, which might negatively impact accuracy. More advanced techniques (hog features, stroke sequence...) would result in improved performance.

• *** Preprocessing does not scale image with stroke width, which leads to poorer results as the image size increases.

Popup screen	Prediction result

Installation

1. Install Python >= 3.6
2. Install package manager pip
3. Install virtualenv

$ pip install virtualenv

4. Create an environment

$ virtualenv ENV

5. Activate the environment

(Posix)
$ source /path/to/ENV/bin/activate

(Windows)
$ \path\to\ENV\Scripts\activate

6. Install requirements.txt

$ /path/to/ENV/bin/pip install -r requirements.txt

Usage

1. Run run.py in root

$ python run.py

2. Open localhost port server link

See info.html for further details.

Contributors

• josugoar - Main contributor - GitHub