TimeandPlace - AI Outfit Analysis

Time and Place uses real-time computer vision and a multimodal AI model to help users decide if their outfit fits both the event’s dress code and the environment.

Comment

Inspiration

What inspired me to create TimeandPlace was hearing stories of people showing up to events either underdress or overdressed. I wanted to create something to help others become more aware in that regard.

What it does

It takes user input (dress code and temperature in Fahrenheit) and analyzes webcam feedback to analyzes the users outfit. Then, the user is given feedback based on their input on whether or nor their outfit is appropriate for the occasion.

How we built it

I first integrated CLIP, a model by OpenAi that matches visuals based on given descriptions. I made use of this aspect of the model by making prompts for each of the four dress codes. I programed in different responses to the model's analysis . For the weather aspect, I integrated a range of temperatures that would be appropriate for each dress code that would also be taken into account by CLIP. I also installed Django so the python logic could be in the web app.

Challenges we ran into

The main challenges for me were frontend-related specifically, CSS. I had trouble with getting the elements on the web page to

Accomplishments that we're proud of

I am proud of myself for successfully integrating machine learning logic into web development.

What we learned

I learned how to use python in web development with Django.

What's next for TimeandPlace - AI Outfit Analysis

I want to integrate more in depth feedback as well as make a more polished UI.

Share this project:

Updates