Inspiration
I have been learning Machine Learning for a couple of months but I haven`t created any final projects. I have been learning python and machine learning algorithms. When I saw Bay-Area-Hacks with the theme of Machine Learning, I felt like this is an opportunity to learn and grow my skills. So, I tried to give my best to the hackathon project as I could give.
What it does
My project is a Web App, It takes images from the users that have text, and only shows the text that is in the images using Artificial Intelligence and Machine Learning Service from Azure. It can be used to digitize any hand-written, or picture receipts/bills. For example: If some students have hand-written notes, or pictures of whiteboard notes from teacher taken on class, then they can easily convert and extract text from these images using my webapp.
A person taking notes on paper at the shop can digitize his whole bills/account with the help of this app and it would be really time-saving to people.
How I built it
I built it using Python-Flask and using the cognitive Machine Learning services provided by Microsoft Azure.
Challenges I ran into
I faced really a difficulty in uploading my site to Heroku. It has created many problems and still, I don't know if my app will be deployed but I m writing this side by side. I completed my project locally but it is not running on Heroku. And Only half-hour is left now. So I am really afraid about project and my hard work too.
Accomplishments that I'm proud of
I am proud that I successfully built and useful machine learning application and deployed it too. Literally, I am really proud.
What I learned
I learned the CLI commands to deploy the flask app to Heroku and learned to use machine learning APIs and services provided by Microsoft which I had never done before.
What's next for Digi-Text
Due to time limitations, I could not do all the features that I wanted to do but I am sure I can add a Language translate feature within my app so anyone can translate the images in any language and read them out from the images and be able to copy the extracted text too.
Log in or sign up for Devpost to join the conversation.