Fashion Lens

Banner
Community Page
Tank Top
Biker
Login
Hoodie and Sweatpants
Jump Suit
Tanktop
Biker Clothes
Character Rigging
Character Rigging
Character Rigging

Inspiration

Finding the right clothes with respect to size and style can be challenging. With technologies like the metaverse and augmented reality (AR) headsets becoming more mainstream, we believe AR can revolutionize the fashion industry.

What it does

Fashion Lens allows users to try on clothes before buying them via AR. Users can also share results in the community section of our app which showcases the designs of various clothings.

How we built it

The app utilizes Apple's ARKit and MLKit to detect pose estimation and the location of users in order to project a 3D model onto them. The 3D models are created and rigged in Blender, so they can morph to a user's body shape.

We utilized Google's authentication API's to allow users to register and login to our service. Swift was used for the business logic of the application.

Challenges we ran into

There were very few techniques published online for AR clothing. The ones published on GitHub only utilized basic computer vision algorithms and lacked proper depth tracking. Thus, our team utilized their years of professional and hobby experience with AR technology in order to engineer a solution that combines ML with SLAM data from sensors to accurately wrap clothes around users.

Accomplishments that we're proud of

We were able to create an AR clothing system that performs better than currently published open source projects.

What we learned

We learned how to create applications using XCode and Swift along with using Blender to modify 3D models.

What's next for Fashion Lens

A link to our application can be present in major retail apps such as Amazon and Nike. This would allow users for these apps to easily try on clothes without worrying about how they would fit or look. Users posting their AR clothing results in the community tab of our app would also allow others to discover new fashion styles.