Inspiration

Whenever we went shopping, we were overwhelmed by the variety. Our tool allows businesses to help uncultured people like us figure out what we should buy and wear, based on designer color palettes that the user can browse and choose from.

What it does

Fresca generates color-coordinated outfits from extensive clothing catalogs. Our tool matches clothes with an aesthetically pleasing color palette to generate personalized outfits for the user.

How we built it

  • Began with accumulating Uniqlo catalog pictures
  • Wrote a script to write prominent color RGB values from catalog pictures to a CSV
  • Wrote algorithm to match colors from a color palette to items in a catalog, optimized as needed for nice lookin' fits
  • Made a front-end app in Android Studio to showcase the results of our color extracting and matching algorithm in a sleek UI, generating easily scrollable outfits with displayed price
  • A lot of “what the f-” in-between

Challenges we ran into

  • Parsing background colors from clothing colors
  • Comparing clothing color to palette color
  • Getting the color algorithm to generate more diverse outfit sets that better represented the chosen palette
  • Literally anything in android studio, it never does what you think it should do or what you want it to do :(
  • The app works seamlessly in the emulator but runs into issues on some android devices. We are still in the process of figuring out why so fair warning if you decide to try to the apk :)

Accomplishments that we're proud of

  • It kinda works :)
  • Created a mostly successful color matching algorithm
  • Expedited our workflow with various data processing techniques that just made our lives easier
  • The palette color names

What we learned

  • How to apply data science techniques to sort through clothing catalogs
  • How to compare colors through numeric values
  • Became proficient an Android studio
  • The process of designing an algorithm can be kind of rough

What's next for Fresca

  • Currently, the app works seamlessly in the emulator but runs into issues on some android devices, and it would be pretty epic if it worked on all those devices
  • Enhance the user interface with swipes and dropdowns and more wardrobes
  • Create a palette generator to generate more palette options to choose from upon scroll
  • Improving consumer experience by allowing users to insert their own wardrobe and custom palette into a custom database for the app to choose a fit for the user
  • Display multiple versions of the same chosen wardrobe from multiple retailers to allow consumers to make economic choices for their

Built With

Share this project:

Updates