FastFit AI

• 

  Generated outfits

• 

  Classification Model

• 

  Python Server

• 

  Generate Outfit

• 

  User Garments

Inspiration

We identified a great opportunity to use computer vision and machine learning to reduce unnecesary time all people spend on a daily basis by choosing what they would wear.

What it does

The user uploads an image of a garment (top, bottom o shoe), our ML model determines the category of the clothing and maps the parameters of each garment (weather and formality). Then another ML model generates multiple outfit suggestions according to the ocassion the user need it for.

How we built it

We used the following technologies to build the platform:

• Neural Networks (CNN)
• PyTorch, Keras, CUDA
• Computer Vision and Machine Learning
• Frontend: Angular; Backend: NodeExpress
• Data Backend: Python
• DB Manager: MySQL
• UI/UX Design: Adobe XD

Challenges we ran into

We ran into the following challenges:

• Hardware limitation (GPU)

Accomplishments that we're proud of

• The model has an accuracy above 85%

What we learned

• We learned to create a local LAN to connect each PC to a local DB; also we mastered ML and CNN for computer vision.

What's next for FastFit AI

• Infrastructure integration with Cloud Services
• Model Optimization with bigger batch sizes (actually limited by hardware)
• Business model integration to social functions
• Web Backend connection with data backend
• Fuzzy Logic for garment color extraction and combination

Built With

• adobe
• angular.js
• cnn
• cuda
• keras
• mysql
• node.js
• python
• pytorch