FACE AND EYE BLINK RECOGNITION USING FEDERATED LEARNING WITH OPENCV2

Abstract

This research provides a Federated Learning Process-based technique for
blinking-based liveness identification of human face. All you need to record video clips using our
technology is a web camera. A blinking clue constitutes a passively action; it doesn't need the
individual performing it to provide any other cues, like speaking or moving their face. We use
CRFs to simulate blinking behavior, which allows for the incorporation of long-range contextual
dependencies inside the observation sequence. The experimental findings show that the
suggested approach works better than the federated learning Processes technique, indicating its
promise.