Archives

Unconstrained Joint Face Detection and Recognition in Video Surveillance System


Nariman Jabbar Qasim and Israa Al_Barazanchi
Abstract

The advancement in computer vision technology and the changes in image and video acquisition techniques such as smartphones, IP cameras, and surveillance systems have made video processing an important field of research. The primary application of video processing has been a new feature in forensic science and law enforcement. The recent applications in video processing are based on human-computer interaction, behavioral analysis, facial and Iris recognition using various machine learning technique. Facial recognition systems, especially in video sequences, has become a fundamental component of most industries today. The time utilized for implementing algorithms associated with facial recognition has been limiting long. In this study, we propose an advanced mathematical logic that will significantly reduce the time for facial recognition and classification. The proposed fuzzy logic approach employs a combination of modeling technique that can easy identify and classify a human face within a video. This technique is implemented in MATLAB, and its primary advantage is that it does not affect the quality of the video all the facial image during the detection time. This technique utilizes two steps in the creation of a facial square in which the primary features are located in the square. The two methods applied other detection of the facial texture and the implementation of geometry to detect the face. The results illustrated in MATLAB demonstrate the success of this fuzzy Logic technique in detecting facial structures in the video.

Volume 11 | 01-Special Issue

Pages: 1855-1862