Have you ever been grocery shopping and wanted a quick and easy way to record and compare prices? GrocerBase lets you quick record that information with just a picture and a little drawing and let's you compare to the other entries made by other users and yourself easily.
What it does
It allows one to record the price, name and other information about a product by simply taking a picture and highlighting important info. Then, our program crops the image to contain only the highlights and uses google-cloud's vision API to read and identify the text and returns that info. Then, both the pictures and the text are stored so that other users may have access to the same information.
How I built it
We have multiple versions (none of which are finished). We have a web based version that makes use of Flask and React and one which makes use of Kivy. These are features on the app and web branches of the github repo. In all branches of the repo, there is a folder called ReadImage Development Folder which has our google API code and the cropping functionality.
Challenges I ran into
Many, the team working on the web app had a very difficult time making react and flask work together. The group working with kivy had no prior experience with kivy, so they had to learn how to write code in it entirely during the hackathon. One of the specific issues we faced was that we needed a separate version of openCV for android and desktop, and the android version was deprecated, so it was hard to find. We also had alot of difficulty getting the coordinate of the highlighting relative to the image, as kivy is designed to be completely resizable. This meant that the image was not displayed at full size and we could not find a way to get the actual size.
Accomplishments that I'm proud of
We over came many issues. Even though we could not finish either version, both have a fully functioning camera capability. The mobile app version also has the ability to highlight the image, but it is still buggy.
What I learned
We learned alot about combinging frameworks from trying to make React and Flask work together and became competent kivy developers in only 36 hours.
What's next for GrocerBase
Completing the code. We need to first get the highlighting working so we can use the google API. After that, we will need to get the server done so that we can store data and have users access it. After that we need to implement the searching function and then we need to fix the stylistics of both versions.