As students we are often faced with the challenge of doing our washing and doing it correctly. After too many a time shrinking a lovely pair of jeans by having it on the wrong setting, we decided that enough was enough and we wanted to develop an app that could easily and automatically take an image of a care tag and tell us what it meant.
What it does
It's a web app (and in the future with more time, potentially a mobile app) that takes an upload of a single image of a care tag and identifies it, also returning a description of what the symbol means.
We were planning on using Twillio to implement a mobile version of the application where a user could send a picture of a care tag to a WhatsApp number and have it return the details. We got relatively far with this, but couldn't complete it in the time given as our team member with the most experience in this had to leave early.
How we built it
We used a Django web framework to build the website. We use an sqlite database in the form of a data model built with Django to have a database of all possible care tags, their images and descriptions. We used machine learning alongside a tensorflow model we trained and customised to use image identification to identify the names of each of the symbols.
Challenges we ran into
Originally, we wanted to take a whole label and identify each of the individual symbols on it. However, splitting the label up into several smaller pictures proved to be a difficult and ambitious task in the time frame, so we instead used single cropped images as inputs instead. We also wanted it to work with a WhatsApp API that would send a care tag as a text image and return a description, and we got quite far with this but couldn't finish it in the time given. We will continue to work on this in time.
Accomplishments that we're proud of
We are proud that Theo was able to get the complex machine learning models to work in such a short frame of time. We're proud that Tara build the functionality of the web framework, Kaal was able to build the backend database and functionality and Lyla helped everyone individually and built the front end so well. We're proud for Henry for being help so much with the machine learning and going so far with the mobile application.
What we learned
We learned so many new skills. We work so well as a team and we learned communication skills. We learned tons of new practical skills, such as Django, building Tensorflow models, how to handle HTML and CSS, handling and querying database structures.
Log in or sign up for Devpost to join the conversation.