Facial Recognition

Facial Recognition

Name

Institution

Abstract

With the extensive use of video surveillance cameras and mobile phones equipped with high-resolution cameras, the increase in video footage or still photographic evidence has been on the rise. Law enforcement officers have the liberty to use these footages and photographs to identify the potential persons of interest. By the use of various developed algorithms, the law enforcing agents can match faces of persons they are interested in with the information found in their database. However, this method of identifying suspects has proven to be less reliable. This is because of the many limitations in capturing clear footage or a still photo of victims or suspects. This paper is going to give an overview of some of the scenarios that facial recognition technology might fail to produce the desired results.

Introduction

Human beings have always been known to use faces when it comes to identifying or recognizing individuals. Over the past years, advancement in computing technology has enabled similar recognitions to be done automatically.  Significant progress in this type of technology has been witnessed in the past ten to fifteen years. This technology is used for both identification and verification. Face recognition is a relatively new concept that was first developed in the 1960’s. According to Goldein, Harmon and Lesk (1971), the first semi-automated facial recognition system was required to locate facial features such as the nose, ears, mouth and eyes in a photograph prior to calculating distances and ratios to a standard reference point. There are two methods of facial verification approach; these are facial matching and facial representation. Hua and Akbarzadeh (2009) argues that description based methods have been proved to be effective face images since they produce the best results. However, certain existing handcrafted encodings suffer two disadvantages. Getting an optimal encoding method manually is difficult. Secondly, handcrafted codes are usually not evenly distributed. Some codes may or may not fail to appear in the final face images. The following study is going to take a look at how facial recognition technology works in the fight against crime.

How Facial Recognition Technology Works

The facial recognition technology uses software that has the capability of picking a person’s face from the crowd and comparing it to a database of images that are already stored. The software must have the ability to differentiate between a primary face and the background. The software used in facial recognition is based on its capability to recognize a face and give various measurements for various facial features. Each and every face has some distinguishable characteristics and landmarks. The different valleys and peaks found in an individual’s face make up the facial characteristics of a person. These markers are known as nodal points, and an individual has approximately eighty nodal points. The nodal points measurable by the facial recognition software are: