Inspiration

We are inspired by 3D spinning product marketing ads, and we wanted to bring this (expensive) look to store owners, artists, and small businesses everywhere at a price that is within reach. We also hear that 3D/360s boosts sales by quite a lot, and it's just plain cool!

What it does

The user takes a video of their product in the app, either by walking around it, or by filming it on a turntable. After uploading the video, the app processes the frames in 3D on our ai cloud platform. After about 15 minutes, the user will have a spinnable 3D product view that they can share with others, or embed on their website.

How we built it

We use the latest research in Radiance Field technologies to build a proprietary machine learning trainer that generates a 3D model from photos. We utilize three.js and webgl to build the viewer. Everything is hosted on the cloud using cloudflare and global cdns. For the app we utilize cross platform Expo and react native, and RevenueCat SDK for in-app purchases.

Challenges we ran into

Sometimes the 3D trainer doesn't capture materials well (highly reflective), so we have an option to use the 360 photos as well for things like Jewelry and Luxury items.

Accomplishments that we're proud of

We're really proud of the 3D Math involved along with the training improvements we've made, and how quickly you can get from object to 3D now!

What we learned

We learned a lot in how to make a multi-platform app (desktop web and mobile ui), along with our learnings from optimizing the app and viewer for mobile specifically.

What's next for Optimal 360

We're improving the backend to work better with outdoor captures in the next month (with higher detail!). We anticipate adding impressive animations that look like expensive advertisements. We also plan on adding instagram and tiktok export for sellers to utilize the app for social media marketing. We plan on adding an explore feature and subscription tier for better art and product discovery, allowing for publishing on the platform. These features should be ready in the next couple months.

Share this project:

Updates