Facial images are commonly used and humans can associate easily with facial images meaning the facial features can be integrated into the technological platforms (Rua et al. 2012). In addition, the facial details can be collected from a distance while voice information can be collected and can be associated with behavioral characteristics of an individual (Giot, El-Abed & Rosenberger, 2013). Hence numerous sources of biometric information exist, and the usefulness of any modality depends on the economic and ergonomic expectations.
Even though biometric samples can be physically bonded to use, consistent, permanent and unique, the algorithms and sensors used to acquire and later, analyze the information may be imperfect. Sensors are able to introduce electrical and optical distortion (Emmanuel et al. 2014). Some of the information may be lost while others may be distorted. Machines are very good in fulfilling the requirements of security especially retrieving the information, and the success of the biometric systems depends on the infrastructure in place, and strategies employed to collect and analyze the information.
Biometric Processes Biometric technologies and associated systems rely on numerous discrete processes. The discrete processes include template comparison, template extraction, live capture and enrollment. The aim of enrollment is to collect and store biometric samples, which will be used for future comparisons. The storage of archiving system is to store the information or enabling updating the system based on changing variables. The samples collected should be of high quality and consistent in nature to improve the matching performance capability (Chowhan & Shinde, 2008).
Live capture is the ability to capture information from an individual based on the archived information enabling identification of the individual. Template extraction analyses the biometric samples resulting in yielding numerical template. The templates generated are stored for future use, and also improve the speed in which the information can be retrieved (Emmanuel et al. 2014). Algorithmic comparisons are employed in comparing the two biometric templates to make appropriate comparisons. It is important to note that the algorithms and methods employed by these different biometric systems are proprietary in nature; hence, different in nature but shares the same ideology.
Therefore, it is difficult to mix the systems for the purpose of identification. The following biometric process defines the biometric approach in fulfillment of identification requirements. Biometric System Accuracy Testing The importance of the biometric system or the efficiency is premised on its capability to accomplish the assigned duties of identification. In utilization of the biometric system, it may result in false or true match against the biometric sample (Chowhan & Shinde, 2008).
The false match rate is defined as the frequency in which the information assessed from different sources are assessed as from the same source while false non match rate shows a sample from the same source but in real sense, the information is assessed from different sources (Henniger & Nikolov, 2013). An effective biometric system should have low rates of false non-matches and false matches, and the results are prompt in nature. In addition, the higher availability of information improves the efficiency of the information because there are basis in which the information can be compared (Phua et al. 2008). It is also important to create a strong system that shields the system from readily available information e.g. information from public systems.
Hence the uniqueness and volume of the information including quality improves the efficiency of the biometric system accuracy. Biometric Applications The use of biometric system is traced to its use in identification of a suspect in a criminal investigation. With the advancement of powerful computing and image capture technologies; the strategy has changed to highly automated and digitalized system compared with previous processes associated with labor intensive and paper based (Chowhan & Shinde, 2008).
Read More