An Automated Scheme For Surveillance Videos Footage And The System Searches

The research presented an automated scheme for surveillance videos footage and the system searches Person of Interest (POI) and retrieves video shots for specific POI. An automated system has four stages; 1) to detect shot boundaries and divides original video into shots is called shot detection stage, 2) detection and recognition of face stage that detect faces in video footage using based on Haar-linke features and using the Principal Component Analysis algorithm (PCA) for extraction of features in order to compute and select the eigenvectors with the major eigenvalues, 3) groups and apply face annotation with similar facial features into the same cluster, and 4) retrieve specifics person footage from database user generate query by person image. The proposed scheme is user friendly and simple. It really overcome workload and enhances accuracy to localized person of interest in surveillance videos data. The average recognition accuracy on NRC-IIT [1] face video data set is 83.2 percent.