PicNote

www.picnote.tech is a website that allows you to turn pictures of lecture slides, notes, textbook pages, and homework into an editable Word document

Comment

picnote.tech

Inspiration

Copying down notes from lectures is quite time consuming and, ironically, makes it more difficult to absorb the material being taught. This made us wonder- what if we developed an application to collect these notes automatically, so we can focus on learning rather than notetaking?

What it does

PicNote composes a Word document from elements (text blocks and images) found in the submitted image- notes, lecture slides, textbook pages, etc.

How we built it

We used the Google Cloud Vision API to parse the images uploaded by the users to then create an editable file for the user. We hosted this service on Google Cloud Run in a Docker container. Our backend was created using Django in Python. We received our domain name, picnote.tech, from Domain.com.

Challenges we ran into

We had to familiarize ourselves with Google Cloud services, but it was not too difficult since there was a lot of in-depth documentation. One large hurdle was that we had to create a system to track the relative position of elements in the image after they are detected by the Vision API.

Accomplishments that we're proud of

We are very proud of being able to successfully integrate Google Cloud Vision into our website, as well as hosting the website using Google Cloud.

What we learned

We learned that Google Cloud services has numerous APIs which can be used in many different ways. We also became much more familiar with using Google Cloud services and deploying our own website.

What's next for PicNote

We would like to implement our own training dataset for our object detection AI to more accurately detect images on surfaces. Google Cloud Vision works well for text, but differentiating diagrams and hand-drawn objects is more difficult.

Built With

Share this project:

Updates