Inspiration
We know how difficult and overwhelming it can be to figure out your undertones and which colors complement your natural features best. We wanted to create something that feels accessible and fun without being overly technical. Our goal was to make this increasingly popular experience more accessible through technology.
What it does
Loopy analyzes the user's facial photo to determine their seasonal color palette and provides personalized recommendations. It guides users on how to take the right photo, processes the image and delivers clear, visually pleasing results that can be easily understood and downloaded for future reference.
How we built it
Our frontend was build with React, Typescript and Tailwind CSS. Our backend is powered by Python and FastAPI. The AI pipeline consists of three different layers, facial analysis and alignment using the DeepFace library, feature color extractor (Color Extraction model based on the adikuma/Colour-Extract architecture) and lastly seasonal classification model which categorizes the user into a color season.
Challenges we ran into
We initially faced an issue where the color extractor returned monochromatic results (ex: only shades of blue). This was traced back to a bug in our custom adapter's logic. Additionally, we struggled with increasing our accuracy considering we had a limited data available for seasonal matching.
Accomplishments that we're proud of
We're proud of successfully integrating a machine learning model into a user-friendly product. Despite the technical complexity behind the scenes, we managed to deliver a smooth, intuitive experience that feels accessible and engaging.
What we learned
We gained experience implementing a multistage AI pipeline. We learned how to use our time efficiently and how to rely on our teammates. We also learned how to add animations to our frontend to make the website more engaging.
What's next for Loopy
Next, we want to increase the data in order to increase the accuracy of the season classifier. We also plan to explore mobile optimization and accessibility improvements so Loopy can reach a wider audience and become an interactive personal companion.
Log in or sign up for Devpost to join the conversation.