Inspiration
We solve the following problem: help people to understand objective feedback on how their styles look -- by community, themselves, professional stylists, and general rules. Targeting market is around 3 billions $.
What it does
Product Concept
- Person takes photo.
- We evaluate color and proportional characteristic, and provide automatic feedback.
- With sentiment analysis (images still recording), we understand if person likes her look.
- The look can be shared for estimation of expert/community.
How we built it
Style checking is done this way:
- Selenium (download colors).
- Deep Learning CNN for types.
- Deep Segmentation, remove background.
- Color neighboring.
Sentiment analysis is approached via Google/Azure API.
Frontend/Backend are done with Android Studio and Flask.
Challenges we ran into
Making working prototype in short time.
Accomplishments that we're proud of
We completed 60% of planned.
What we learned
For such hackathons, preparation on Android Studio sides should have been made.
What's next for Cherry Fashion
Looking for customers and interested industries.
