Multiple Face Detection Tracking and Recognition from Video Sequence

Abstract

AbstractFace detection and recognition are a cutting-edge biometric technology. Numerous approaches and systems have been extensively studied in this subject. Face recognition is gaining popularity, and the majority of us use it without recognizing it. This detects many faces from a video sequence that has been tracked and identified. Face detection is performed here using the Viola–Jones technique and neural networking. By integrating the Viola–Jones approach with a neural network, it is possible to determine the computing time and develop a resilient algorithm (this system will adapt and learn new data very fast and system more robust). The system becomes more robust and capable of detecting more faces, while reducing false positives. Neural networking is a method for detecting faces that is image-based. For image tracking, the KLT tracking algorithm is employed, and for recognition, the eigenface detection approach is used. The purpose of this study is to detect many faces in a video sequence, follow the face and recognize it.KeywordsKLT trackingViola–Jones algorithmEigenface methodHaar featuresNeural networkMultiple face recognitionMultiple face detectionMultiple face trackingRegion of interest (ROI)