TrueFace
Safeguarding Authenticity in Media
The problem TrueFace solves
AI based technology has evolved to the point of creating almost indistinct fake videos, photos and articles. Apart from spreading fake news and propaganda, this has now being used as a tool for various security and malware attacks. On top of that, AI generated media is becoming increasingly sophisticated, making it difficult for the average internet user to discern real from fake. Our solution - Trueface, is a website extension with real time deepfake detection and fake image and news detection. We believe everyone deserves to browse the web with confidence and avoid falling victim to manipulation. Our extension integrates seamlessly into users' browsing experiences. No complicated setup or technical expertise is required, just install and start browsing with confidence.
By providing real-time deepfake detection, along with AI generated image and text detection, our extension aims to be a valuable tool for anyone who wants to navigate the online world safely and make informed decisions about the content they consume
Challenges we ran into
One of the main hurdles we ran into was converting our Machine Learning model into an ONNX file. ONNX is basically an open-source format that facilitates interoperability between different deep learning algorithms for simple model sharing and deployment. We solved this issue by modifying the input and output parameters, and taking the guidance of the open source documentation available.
Another issue we faced was the lack of resources pertaining to running inference of the ML model in the chrome extension. We opted to keep things on the client side, without using the server, keeping the privacy of the users as our main priority. We resolved this by extensively reading through the documentation in the limited time provided, and tried our best to integrate the model with the extension.
Tracks Applied (1)