Abstract
Biometric based attributes are the latest additions to the existing mechanisms used for security of information system and for access control. Among a host of others, face recognition is the most effective biometric system for identification and verification of persons. Face recognition from video has gained attention due to its popularity and ease of use with security systems based on vision and surveillance systems. The automated video based face recognition system provides a huge assortment of challenges as it is necessary to perform facial verification under different viewing conditions. Face recognition in video continues to attract lot of attention from researchers world over hence considerable advances are being recorded in this area. The aim of this chapter is to perform a review of the basic methods used for such techniques and finding the emerging trends of the research in this area. The primary focus is to summarize some well-known methods of face recognition in video sequences for application in biometric security and enumerate the emerging trends.
TopIntroduction
Today, in the era of Information Technology (IT), security is a major concern of every individual, organization and even nations all over the world. Security and surveillance, identification and authentication methods have been challenging issues in various areas like entrance control in buildings, access control for computers in general or for automatic teller machines (ATM) in particular, day-to-day affairs like withdrawing money from a bank account or dealing with the post office; or in the prominent field of criminal investigation.
Different security mechanisms are required and extensively deployed in such areas where shared accesses are applied as part of a composite mechanism. User authentication is a major challenge for security of any system. Though passwords and PINs are used to grant access to information system, they are hard to remember and can be stolen or guessed. For shared accesses, though cards, tokens, keys etc. are used, there is the possibility of misplacing, forgetting, stealing or duplicating these items and hence they are not much reliable. Even magnetic cards are also not reliable as they may be corrupted and become unreadable. In comparison to these, an individual’s biological traits are more reliable means to provide accesses to secure a system as they cannot be misplaced, manipulated, forgotten, stolen or forged. Since security is becoming major concern to provide access to sensitive information and locations, such biometric based systems are becoming most popular today.
Biometrics refers to measurement attributes related to certain parts of the human body and the related characteristics. Biometric technology is defined as automated methods of verifying or recognizing the identity of a living person using his/her physical or behavioral characteristic (Miller, 1988; Wayman, 2000). The term “Biometrics” is the abbreviated term for “Biometric Technology” that recognizes people as individuals with the help of computer (Wayman, Maltoni & Maio, 2005). Both physiological and behavioral components are used as the measures for identifying individuals using biometrics. Physiological characteristics are related to the shape of the body. These include faces, fingerprints, finger geometry, hand veins, palm, iris, retina, ear, voice etc. On the other hand behavioral characteristics are related to the pattern of behavior of a person. These include signature, writing style, keystroke dynamics etc. (Jafri & Arabnia, 2009). These characteristics vary from person to person. Therefore, biometrics can be used to uniquely identify or verify a person to grant access to any logical or physical system based on his physiological and behavioral characteristics. Besides uniqueness property, another advantage of using biometrics is that it is always available with the person.
Face recognition is a biometric method used to verify and identify an individual automatically through a computer application. Face images are most common biometric characteristic used by humans for identifying persons (Jain, Hong & Pankanti, 2000). Face recognition approach is based on shape and location of the face attributes viz. eyes, eyebrows, nose, lips and chin shape and their spatial relationships (Ross & Jain, 2007). Among the biometric methods, face recognition has several advantages as face can be recognized passively without any explicit action or participation from the user end (Jafri & Arabnia, 2009). For other biometric systems like fingerprint, retinal or iris scans, the user has to perform an action similar to scanning his finger voluntarily for his authentication, verification and identification. Thus these methods rely on the cooperation of the participants. On the other hand, for machine aided face recognition, the user is automatically identified by simply walking through a surveillance camera, without the participant’s cooperation or knowledge. Even facial images can be easily obtained with a certain numbers of inexpensive digital cameras. However, for surveillance systems, video cameras are essential and the feeds come continuously. Human beings often recognize one another by unique facial characteristics. From the literature it is ascertained that face recognition is the most successful form of human surveillance.
Key Terms in this Chapter
Face Detection: It is process to determine whether or not there are any faces in an image or video and if present, return the image location and extent of each face. The output of the face detection system is an image region containing only the face area.
Feature Extraction: It is a process for extracting relevant information from an image. After detecting a face, some valuable information are extracted from the image which are used in next step for identifying the image.
Verification: It is the process of establishing the truth, accuracy, or validity of something. For person verification, biometrics are used to verify an individual besides using some traditional methods. Image based person verification is a system which is a one-to-one matching to verify one individual’s claim that who he is by comparing his image with the images already stored in the database.
Face Recognition: Face recognition is a biometric method used to verify and identify an individual automatically through a computer application.
Face Tracking: This process accumulates both spatial and motion information in subsequent frames in a video and finds image regions for faces detected in previous frames. Both detection and tracking search for features in the image that indicate the presence of face.
Authentication: It is the act of confirming the truth of an attribute of a single piece of data claimed true by an entity. It is a process in which the credentials (e.g. username and password) provided are compared to those in a database of authorized users' information on a local operating system or within an authentication server. If the credentials match, the process is completed and the user is granted authorization for access.
Person Identification: It is a process that identifies an individual uniquely and permits another person to assume that individual’s identity without their knowledge or consent. An individual can be uniquely identified by using unique personal identification number (PIN) and also using biometrics like fingerprint, face etc. Image based identification system is a one-to-many matching to determine an unknown person’s identity by comparing his image with the images already stored in the database.
Biometric Security: It refers to authentication techniques that rely on measurable individual’s physical characteristics that can be automatically checked. The system measures and analyzes human body characteristics, such as DNA, fingerprints, eye retinas and irises, voice patterns, facial patterns and hand measurements, for authentication purposes. Biometric security systems enhance security by linking a unique physical attribute of a user to the data that they are allowed to access.