Handwritten To Text Output (HTTO)

Authors: Armando Cabrera & Jaime Andrade

This is a still-in-work prototype for the course CS4361 at The University of Texas at El Paso.

Required Libraries

• pip install Flask
• pip install tensorflow
• pip install numpy
• pip install Image
• pip install BytesIO
• pip install jsonify

Download files from Repository

• server_test -> app.py
• server_test -> templates -> index.html
• server_test -> static -> script.js, style.css
• models -> handwritten_character_model_enhanced.h5

(OPTIONAL)

• handwritten_test_images
• dataset

User Interface

Once you have downloaded all required files, in your preferred coding environment run this command in the terminal:

• python .\server_test\app.py

You will be prompted with tensorflow activation routine, amongst in there will be a line:

• 'Running on http://127.0.0.1:0000'

Cntrl + Click on that link

You will then be prompted with a screen similar to the one below

Intuitive from Now on!

• Simply just upload the desired image you wish the model to 'attempt' to read
• Click on 'Upload'
• and BOOM, will probably be the incorrect output but will eventually get it right.