SmartFit

A website that identifies clothes a person is wearing. User uploads an image of a person wearing an outfit, and site uses machine learning to classify the different pieces of clothing in the image.

Comment

Inspiration

One of our group members Kevin Chong has a very bad fashion sense and does not know the names of any articles of clothing. The other member Albert Kang came up with the genius idea to create an app that helps Kevin and others similar to him with this issue. We wanted a simple way for people to be able to expand their clothing knowledge, while also checking other people's fits. Thus, SmartFit was born

What it does

It prompts the user to either select or drag an image into the input space of a person wearing an outfit. Then the website runs the image through our model to determine the top and bottom pieces of clothing that is in the image. Then the website will display back to the user both the name of the top and bottom piece of clothing the person in the image is wearing. The model we use can detect 11 different types of tops and 7 different types of bottoms.

How we built it

Loaded our dataset of clothes and then used Keras to generate our machine learning model. After training the model we hosted a new website locally using Node.js. We then used Django to run all the python scripts we had made. Then created the UI and styled our website.

Challenges we ran into

Training the model to be accurate Getting the image upload feature to work

Accomplishments that we're proud of

Getting python code to work on a website Creating and training a machine learning model

What we learned

How to use Django Some HTML and CSS tricks

What's next for SmartFit

Suggest fits for users to wear Provide more information on the outfits in the image (color, price, etc.)

Share this project:

Updates