Envision

Elevating the consumer's online shopping experience through increased visible accessibility.

Comment

Inspiration

As young students who love to thrift and shop online for convenience, there have been numerous instances where the hesitation to follow through with buying the items in our cart peaks. This derives from shopping through apps and websites like Ebay and Depop, where resellers who list their items with sizes that are hard to imagine worn. As such, online shopping is a gamble of trust between the consumer and website, which results in no purchases. We wanted to change that relationship to not only increase the rate of purchasing, but also boosts personal physical confidence for every online shopper.

What it does

Our website guides the consumer to design their outfit using clothing items from their shopping carts, or already owned items from their closets, to then be applied to a provided image of themselves posing, so that the final reveal will entail the new, fully clothed consumer. Additional sizing preferences will be available for input depending on the type of fit the user would like, including tight, normal, and loose.

How we built it

Utilized AI diffusion models with segmented masking.

Challenges we ran into

Masking, pre-trained model conflicts, and clothing sizing with respect to the human model.

Accomplishments that we're proud of

Satisfied with the quality of the output of the diffusion models and the overall appearance and usability of the website.

What we learned

I should not be as nervous about AI stealing my job, but rather the impact it has as a tool.

What's next for Envision

Further applications for our website we would want to pursue include more accessibility for sizing, a more in-depth trained model with more clothing item applications, an app that can upload the photos directly, and additional features for the closet. Most importantly, our version is only ran locally, so we would want it to be deployed.

Built With

Share this project:

Updates