PhotoTUM

Organize your event, collect and share your best moments, all in one place.

Comment

Inspiration

Even the best organized events often lead to confusion on the participants side. Too many different platforms, mails written, discord servers joined, and passwords created... To streamline the process, we asked ourselves, what if we could merge all of these processes into one...

What it does

We prototyped an application that lets you chat with people, check out announcements, upload pictures and use facial recognition to find all the pictures you are a part of, create memes inside the app and share them with others, create ToDo's and handle administrative tasks related to registration.

How we built it

We used React Expo to built a cross platform protoype and provide an impression of the UI/UX. For the facial recognition functionality, we set up a local node.js server that runs the facial_recognition library, a python wrapper for the open source C++ dlib library (1. face detection -> HOG (Histogram of Oriented Gradients); 2.Face landmark prediction -> trained shape predictor model; 3. Face embedding -> 128-d vector)). other: Python, CSS, HTML

Challenges we ran into

We had a slightly late start therefore ran into a bit of timepressure. Working in parallel on facial recognition and the ui/ux of the app, we had some trouble merging the facial recognition functionality into the latest version of the app prototype. Also creating an intuitive design for the wide range of components we wanted to add wasnt easy. Trying to work with the api for TUMs building plans didnt work out for us, instead we went for open street maps to give an idea of what could be.

Accomplishments that we're proud of

We created a really cool design, that looks like a fully production ready application, through many iterations we were able to create a great UX throughout the app! We also managed to set up the facial recognition functionality and were able to use it successfully, all of which was hosted locally, without external services such as AWS Rekognition api. Even though we had a late start, we managed to create something we had fun building and we think could be useful and something people will like!

What we learned

You can do a lot more than you think when you put your mind to it. Putting yourself in a situation such as HackaTUM where there is no alternative than to keep pushing, really shows you how much you can achieve. Speaking practically, we learned a lot about the difficulties of building a thought through UI, and learning new technologies, in this case, facial recognition.

What's next for PhotoTUM

We would love to continue building on what we accomplished here, but with a more than full time job and masters studies + work for the two of us, we will see how soon we will be able to return to the project.

Built With

  • expoframework
  • face-recognition
  • react
  • tailwindcss
Share this project:

Updates