Inspiration
We were inspired by the growing risk of visual data leakage and the need for a solution that empowers users to control their own privacy, rather than relying on third-party services.
What it does
Our prototype uses on-device Generative AI to automatically detect and obscure sensitive information in photos, such as faces, documents, and license plates. It allows users to view a blurred image by default and unlock the original only with explicit permission, keeping their data secure.
How we built it
We used a privacy-centric framework to build a photo gallery application that processes images locally using a pre-trained AI model to identify and blur specific sensitive regions within the image.
Challenges we ran into
The primary challenges were determining the specific criteria for our AI model to decide what to blur and integrating it with our application's other components.
Accomplishments that we're proud of
We are proud of successfully creating a functional prototype that demonstrates the power of on-device AI for privacy protection.
What we learned
We learned a lot about the nuances of implementing AI models and the importance of a privacy-by-design approach.
What's next for TikTok TechJam Privacy-First Gallery
We plan to expand the app's capabilities to be able to more accurately blur objects in images and better gauge which objects truly carry sensitive data.
Log in or sign up for Devpost to join the conversation.