Face Recognition

Abstract

This research paper represents the ideas about face recognition and, what technology is contributing in face recognition and face detection. There are different methods of face detection which are used through technology and different software. This paper will also cover techniques and usage of face recognition applications in different sectors.

Contents

Introduction4

Materials and Methods5

Detection6

Alignment7

Measurement7

Comparison Matching8

Verification or Identification8

Results and Discussions8

Techniques12

Conclusion12

References14

Face Recognition

Introduction

Face detection is a field of computer vision of detecting a face in a human digital image . This is a specific case of object detection, where it is desired to detect the presence and precise location of one or more faces in an image. This is one of the areas of computer vision among the most studied, with numerous publications, patents, and conferences. Strong research activity in face detection has also led to the emergence of generic methods object detection. Face detection has many direct applications in video surveillance, biometrics, robotics, control of man- machine interface , photography, image indexing and videos, image search by content. It also facilitates full automation of other processes such as; face recognition or facial expression recognition. Face detection is desired to detect the presence and precise location of one or more faces in a digital image (Lienhart et al, 2003). It is a subject that is difficult, especially due to the high variability of appearance of faces in unconstrained conditions.

It helps in doing Intrinsic variability of human faces such as; color, size, and shape, presence or absence of specific characteristics such as; hair, mustache, beard, glasses, facial expressions changing the geometry of the face, occultation by other objects or faces other, orientation and pose face in profile, illumination conditions and image quality. Face detection must cope with a high intra-class variability, and most methods for object detection with rigid objects that are not suitable. Research methods and the first significant develop derived mainly from the 1990s. The increasing power of computers allows the use of statistical methods and learning more complex and larger volumes of data, which allows a net performance gain.

Materials and Methods

It is particularly important because it differs from previous methods of object detection, previously limited to rigid objects, and therefore with less variability. An important step was taken in 2001 with the publication of the Viola and Jones method, the first method that can detect faces in real-time (Schwerdt, 2000). The method becomes standard and improved upon by many researchers. It is also from the 2000s, that appearance based methods, learning models face from a set of training images, generally prove superior to other approaches. The modular technology of detection and identification of persons on two-dimensional images, which includes three main modules

Detection of persons, indexing (coding and quick search) of persons in the database and, identification (verification) of persons Scheme of technology. Modules are applied sequentially and, allocated to the current image frame persons entering the system of indexation, which in response points from 5 to 10 candidates of the stored database of face images, the most similar to the current ...