Find answers to common questions about RecognizeMe AI, our biometric research, and how we handle your privacy and data security.
No. By default, RecognizeMe AI performs on-device processing. This means your facial features are analyzed in your browser's memory and deleted the moment you close the tab. Data is only stored in our secure Firebase backend if you explicitly choose to 'Register' a profile for testing identification features.
No account is required for the standard research demo. We believe in open-access technology, allowing students and researchers to test AI facial landmarks instantly without the friction of a sign-up process.
Yes, RecognizeMe AI is a 100% free resource. It was developed as a student project by Nithish Janapala to demonstrate the capabilities of TensorFlow.js and Face-API.js in a web environment.
Yes. The tool is optimized for modern mobile browsers (Chrome on Android and Safari on iOS). However, because the AI processing happens on your device's CPU/GPU, performance may vary based on your phone's hardware and the quality of the front-facing camera.
Absolutely not. The video stream is processed as a local buffer. We do not have a "record" function on our servers, and we never see your live video. Your privacy is protected by the 'Client-Side' nature of our neural network.
Our models achieve a high confidence score (typically >90%) in well-lit environments. Accuracy can be affected by extreme shadows, low-resolution cameras, or if the face is obscured by heavy masks. We utilize a 68-point landmark mapping system to ensure precision.
If you have registered a face profile and wish to remove it, you can do so through the interface or by contacting our team via the Contact Page. We comply with standard data protection principles, ensuring you have full control over your biometric descriptors.
To perform real-time analysis, the AI needs to 'see' the video frames. The browser permission is a security feature that ensures you are in control of when the camera is active. You can revoke this permission at any time in your browser settings.
The project utilizes a stack consisting of Vanilla JavaScript, TensorFlow.js, and WebAssembly (WASM). This allows us to run complex neural networks directly in the browser without needing a powerful backend server.