Inspiration
Techies can dress well too. Right? The first thing people notice about you is the way you dress. It affects your career, friendships, relationships and even self-esteem. But for everyone style takes time and advice from others to develop. Cue FINE.
What it does
FINE is an automated clothing stylist app that determines if your outfit is FINE or NOT FINE. The user submits a selfie of their outfit which the app uses Fashwell AI to compare to the best fashion runway looks. The app then returns an answer confirming if your outfit is on point or if you should change and try an alternative!
How was the APP built?
Using Android studio, Java was used, with Android SDKs and some useful libraries as Retrofit library and querying the Fashwell API. If the API did not fail us, we will be able to profile user preferences and provide the user with items recommended from the Shiroop shop.
Challenges we ran into
Initially, we were motivated to include machine learning technology to improve the user experience but due to the limit hours and resources, this remained as a proof of concept as a potential future project. Interacting and incorporating the Fashwell visual recognition API with our Android app. Specifically, the API did not recognise the images taken by the camera.
Accomplishments that we are proud of
We have stuck through until the end despite the problems and created a prototype that we believe have unlimited potential once more time is put into it.
What we learned
We absorbed and accumulated experience from different sources and an original idea was born as a result. We also learned to work and cooperate with strangers that soon became a team working in the same direction. The project also gave us a perspective of the challenges involved in trying to integrate an API with Android Studio. Time management was another aspect that we learned which is important for a successful hack.
What’s next for Fine
FINE has multiple extension possibilities that have already been discussed (but limited by the time restrictions). Feature 1 would be to allow the user to personalise their style preference by offering a tinder-esque swiping option to refine the set of best fashion runway looks the app compares your selfies to. Feature 2 is to expand the analysis of outfit beyond 2-pieces (top and trousers) – this may include shoes and multiple upper layers for example. Feature 3 is to offer more detailed recommendations that are either non-binary or suggest alternative clothes to try in particular. Last but not least, Shiroop's API could be included to enhance a user's shopping experience, making it more enriching and memorable.
Log in or sign up for Devpost to join the conversation.