Human face detection in a crowd image based on template matching technique

Abstract

In numerous security applications, face identification and recognition are astounding and significant capacities that we use in our daily lives. The primary purpose behind the interest in building up the computer vision-based robotized advances of mechanized face acknowledgment emerges from the genuine worries for public security in the present organized world. Where personality check's for physical and coherent access in numerous offices are basic in daily life.The process of identifying a human face image through a crowd is an important and very useful task for security purposes. Usually, digital cameras are erected on the roofs of dwellings, buildings, and storefronts to monitor events and accidents on the opposite streets. Often, the captured photos involve a group of people who are at that location. In the event of a strange accident, security services are required to identify the perpetrator. If these security services have pictures of suspects, then we can compare and match their photos with the captured image to identify the real perpetrator. Traditionally, It includes three processing steps in a crowd analysis.That is, preprocessing, object detection, and event/behavior recognitionIn this paper, we will try to recognize a person from a crowd of people by using a single image of the person to be recognized, using the Two-Dimensional Normalized cross-correlation function.