Kunho Kim

Research Scientist [AT] RebuilderAI

kkh23834022 [AT] gmail.com

Fash

Weather based AI style coordination recommendation App (Deep Learning, iOS)

2019 Fall & Winter IT union club Prography, 5th, project

In this work, we develop and deploy the weather based AI style coordination recommendation iOS App called “Fash”. As the team leader, I selected the project item (style coordination recommendation App), designed overall (pipeline & UI), and implement the deep learning (multi-label multi-class classification).
    

“Fash” creates a virtual personal wardrobe based on the images of clothes uploaded by the user and employs a classification model to help categorize styles when adding clothes to the closet. Based on the virtual personal wardrobe, “Fash” provides style recommendations for clothing coordination according to today’s weather. The weather information (lowest/highest temperature of the day, rain or snow) was obtained through Korea Meteorological Administration (기상청) API. Github
    

Poster for Demo Day


    

Main Page of Fash App


    

Prototype of Fash App
For a closer look, please utilize the full screen.