Avatarpose

Inspiration

There's no satisfactory customer value creation process in the current NFT space:

• The collector is mainly left out in the NFT creation process. Even when collectors can co-create, they are limited to generic motion templates ("emotes") or styling options provided by the creator. See projects like Decentraland and Somnia

• The current tools we have for expressing emotion and motion (animations) are advanced, such as Blender and VoxEdit. This leaves out a large creator segment that doesn't have the ability or access to use these tools

What it does

The project allows users to create 3d NFT animations using their real-time body movements from a camera.

Note: For the hackathon, we focus on using hand movements to create handshake animations. This is just one possible use-case. We can also have animations from other body features like your face.

How we built it

Using TensorFlow.js, a hand landmark detection model, and a rigged 3d hand model

Challenges we ran into

We didn't finish the integration of the deployed NFT contract to our frontend. However, I now know how to use the starknetjs library

Accomplishments that we're proud of

• Managed to deploy the NFT contract on Starknet Sepolia testnet

What we learned

How to deploy to Starknet, and how to setup wallet providers on the frontend

What's next for Avatarpose

Bringing 3d handshake animation to Google Meet.

Built With

• cairo
• machine-learning
• tensorflow.js
• three.js