INSPIRATION We got the idea to built this after watching various articles on the deepfake and how some people teases the innocent people. they have to suffer the most . Misleading of information is common now and its hard to find the real one from the naked eyes .
WHAT IT DOES We created the ASA in that way , if we are uploading some fake photo like naked photos or something else , first this will generate the warning if even after that i upload that photo then this model will send the data of me to the server and account can be suspend and they can take some actions on me and that photo will be removed from the social media.
CHALLENGE While working on this project we ran with multiple challenges like debugging the issue , getting proper accuracy. Our projects works on security so it's hard to manage the security of all and after trying multiple times we did it. Variability and Diversity of Deepfakes: Deepfake techniques are constantly evolving, leading to a wide variety of approaches such as face swapping, facial reenactment, and audio-driven synthesis. Detecting all these variations requires a versatile and adaptive detection model.
Real-Time Processing Requirements: Detecting deepfakes in real-time applications (e.g., live streaming, social media monitoring) requires models that are computationally efficient and can process large volumes of data quickly without compromising accuracy.
Ethical Considerations: Balancing the need for effective detection with privacy concerns and potential misuse of detection capabilities raises ethical challenges. Ensuring that detection methods respect individual privacy rights and are used responsibly is essential.
We feel happy that we are contributing to the world with our project and helping to the many people who need this.
We have some plans to implement on it and launch it officialy in no time and we are working on the designing and more accuracy.
Log in or sign up for Devpost to join the conversation.