Source-linked AI summary
Face Recognition Methods & Applications
Divyarajsinh N. Parmar, Brijesh B. Mehta
TL;DR
Face recognition must identify people automatically despite variations in lighting, expression, age, pose, and image transformations. This paper surveys holistic, feature-based, and hybrid methods, illustrates security and other applications, and identifies 2D/3D recognition and large-scale deployment as open research directions.
Problem
Face recognition requires automatic identification despite changes in lighting, expression, ageing, pose, and image transformations.
Method
The paper surveys holistic matching, feature-based extraction, and hybrid face recognition methods, alongside representative applications.
Results
The paper provides an overview of face recognition methods and applications to support understanding of the field.
Takeaways & Limitations
2D and 3D face recognition and large-scale applications remain challenging areas open to further research.
Takeaways & Limitations
Large-scale surveillance remains challenging because varying lighting conditions and face orientations complicate deployment.
Abstract
from arXiv · showhide
Face recognition presents a challenging problem in the field of image analysis and computer vision. The security of information is becoming very significant and difficult. Security cameras are presently common in airports, Offices, University, ATM, Bank and in any locations with a security system. Face recognition is a biometric system used to identify or verify a person from a digital image. Face Recognition system is used in security. Face recognition system should be able to automatically detect a face in an image. This involves extracts its features and then recognize it, regardless of lighting, expression, illumination, ageing, transformations (translate, rotate and scale image) and pose, which is a difficult task. This paper contains three sections. The first section describes the common methods like holistic matching method, feature extraction method and hybrid methods. The second section describes applications with examples and finally third section describes the future research directions of face recognition.
I. INTRODUCTION
Face recognition has developed for nearly 50 years as a pattern-recognition and computer-vision research area with applications including biometrics, information security, access control, law enforcement, smart cards, and surveillance. Recognition systems should also be easy to update and expand to accommodate more subjects.
- I. INTRODUCTION: Face recognition has been studied for almost 50 years and supports applications in biometrics, information security, access control, law enforcement, smart cards, and surveillance.The field is situated within pattern recognition and computer vision.
- I. INTRODUCTION: A face recognition system should be easily updated and enlarged to increase the number of recognizable subjects.
II. Face Recognition Methods
Face recognition methods are organized into holistic, feature-based, and hybrid approaches. These methods range from processing complete face regions to extracting local facial features or combining both with 3D depth information.
- Methods Overview: Face recognition methods comprise holistic matching, feature-based structural methods, and hybrid methods.
- Holistic Matching Methods: Holistic methods process the complete face region, including Eigenfaces, Principal Component Analysis, Linear Discriminant Analysis, and independent component analysis.
- Holistic Matching Methods: The eigenface approach compares an unknown image’s weight with database weights and returns the closest match, while rejecting inputs exceeding a threshold.
- Feature-based (Structural) Methods: Feature-based methods extract local features such as the eyes, nose, and mouth, then classify their locations and geometric or appearance statistics.A major challenge is restoring features hidden by variations such as matching a frontal image with a profile image.
- Hybrid Methods: Hybrid systems combine holistic and feature-extraction methods, generally using 3D images to capture facial curves, depth, and profile information.Depth and an axis of measurement provide information for constructing a full face.
III. Face Recognition Applications
Face recognition is applied across security, identity verification, surveillance, human-computer interaction, and information management. Examples include voter-registration deduplication, computer-terminal monitoring, airport security, and image-database investigations.
- III. Face Recognition Applications: Face recognition supports human-computer interaction, virtual reality, database recovery, multimedia, entertainment, information security, biometrics, law enforcement, and personal security.Applications include operating systems, medical records, online banking, passports, border controls, video surveillance, driver monitoring, and home surveillance.
- III. Face Recognition Applications: Face identification verifies authorized people from face images rather than relying solely on identification numbers, passwords, personal identification numbers, or keys.This establishes the presence of an authorized person through facial evidence.
- III. Face Recognition Applications: Voter-registration systems can compare voters’ face images directly to eliminate duplicate registrations assigned different identification numbers.Highly similar top matches require manual review, and voter photographs may be captured under natural conditions.
- III. Face Recognition Applications: Computer-terminal monitoring can recognize who is present, protect unattended work, and disable the mouse and keyboard after a predetermined absence.A screen saver covers files when the user leaves and normal access resumes when the user returns.
- III. Face Recognition Applications: Airport security systems use face recognition to alert officers when a person resembling a known terrorist suspect enters a security checkpoint.Fresno Yosemite International Airport deployed Viisage technology in October 2001, followed by investigative processing of recognized individuals.
- III. Face Recognition Applications: Large-scale surveillance remains challenging because lighting conditions, face orientations, and other factors reduce user satisfaction and complicate deployment.Other applications include searching image databases and verifying identities for elections, banking, commerce, national IDs, passports, and employee IDs.
IV. Conclusion & Scope for Future Research
The paper presents concepts of face recognition methods and applications to improve understanding of the field. It identifies 2D and 3D recognition and large-scale uses as future challenges.
- Conclusion: The paper introduces face recognition methods and applications to provide readers with a better understanding of the field.It describes the research area as likely to remain active for many years.
- Scope for Future Research: Future research challenges include 2D and 3D face recognition and large-scale applications such as e-commerce, student IDs, digital driver licenses, and national IDs.