Inspiration
We were inspired to create this program because of the recent advancements in ai deepfake videos. Sometimes it is almost impossible to determine if a video is real or not.
What it does
DeepDetech'd scans videos of people's faces and determines if they are deepfaked or if they are real with a trust score on how accurate the decision is.
How we built it
It was programmed almost entirely in python. We used fast api and databases full of deepfaked faces then connected it to a facial scanner software for the backend. We used streamlit api for the frontend website.
Challenges we ran into
We had a lot of issues getting the facial scanner calibrated. At first it was saying every video was fake, then it was saying every video was real. It was very challenging getting the software to be able to accurately detect the authenticity of each video.
Accomplishments that we're proud of
We are very proud of the work it took to get the front end to work with the back end, and how we were able to use multiple apis together for one coherent project.
What's next for DeepDetech'd
We hope to be able to scan photos as well and determine if they are real or not. We also hope to be able to make our scanner more accurate.
Log in or sign up for Devpost to join the conversation.