For verification of this paper, please visit on http://www.socialresearchfoundation.com/innovation.php#8

The aim of the study on biometric authentication is to evaluate and enhance the effectiveness, security, and user experience of biometric systems for verifying and securing individual identities.

Every biometric system consists of four basic modules:

1. Enrolment Unit

The enrolment module registers individuals into the biometric system database. During this phase, a biometric reader scans the individual’s biometric characteristic to produce its digital

Representation.

2. Feature Extraction Unit

This module processes the input sample to generate a compact representation called the template, which is then stored in a central database or a smartcard issued to the individual.

3. Matching Unit

This module compares the current input with the template. If the system performs identity verification, it compares the new characteristics to the user’s master template and produces a score or match value (one to one matching). A system performing identification matches the new characteristics against the master templates of many users resulting in multiple match values (one too many matching).

4. Decision Maker

This module accepts or rejects the user based on a security threshold and matching score

[1]

Classification of Biometric Technology

Two main groups include biometric technology:

1. Physiological Biometric Techniques: These features are innate to each individual's physical traits of the body. Facial recognition, iris recognition, hand geometry, and fingerprints are a few examples.

2. Behavioural Biometric Techniques : These characteristics are linked to unique patterns of behaviour. Voice recognition, keystroke dynamics, and signatures are a few examples. Given that it differs amongst people, voice recognition could also be regarded as physiological.

The major components of a biometric authentication system typically include:

• Sensor: Collects biometric data, such as fingerprints or facial features.
• Feature extraction: Converts raw biometric data into a digital template.
• Database: Stores biometric templates for comparison during authentication.
• Matching algorithm: Compares captured biometric data with stored templates to verify identity.
• Decision system: Determines whether the biometric data matches a stored template and grants access accordingly.

 

Finger Print Technology

Fingerprint identification is the oldest and best-known method of biometric identification for individuals. The important fact is that each person has a different fingerprint. However, the fingerprint identification has been subjected to major changes in recent years. A fingerprint is an impression of the friction ridges of a finger or any part of the finger. A friction ridge is a raised portion of the palmar (palm) or digits (fingers and toes) or plantar (sole) skin, consisting of one or more connected ridge units of friction ridge skin. These ridges are sometimes known as "dermal ridges" or "dermal ". Traditional methods use ink to get the finger print onto a piece of paper. This piece of paper is then scanned using a traditional scanner.[8]

Nowadays, live fingerprint readers based on optical, thermal, silicon, or ultrasonic principles are commonly used.  However optical fingerprint readers are the most prominent.

Fingerprint Classification

Fingerprint classification plays a crucial role in identification systems by reducing the search space within a fingerprint database. By partitioning the database into smaller, more manageable segments, it becomes easier to locate and verify individual fingerprints. This process is particularly useful in systems with low populations, where a single class can be used to identify or verify an individual's identity.

Classification of fingerprints is based on the direction of the ridges around the core, as well as the number of core and delta points. Despite this, it was later discovered that not all fingerprints fit neatly into these predefined categories. Consequently, classification systems often include an "unexpected classification" category for such fingerprints.

Fingerprints are typically divided into five main classifications:

1. Arch

2. Tented Arch

3. Left Loop

4. Right Loop

5. Whorl

Fingerprint Analysis Methods

To optimize database searches, not all fingerprint images are fully stored. Instead, the entire image is initially analysed, and key points or features are extracted and stored. This method significantly speeds up the search process.

Each fingerprint image has around 35 important features, known as minutiae, which include points such as ridge endings, bifurcations (forks), and crossovers. Using between 8 to 22 of these minutiae points is generally sufficient to identify and verify fingerprints with high accuracy.

 

The fingerprint profile, containing these extracted features, can be stored on various mediums, including workstations, servers, or smart cards. This ensures the flexibility and security of fingerprint data management, catering to different system requirements and security protocols.

Face Detection: Special cameras or sensors detect and locate the image of a face, whether it's alone or in a crowd. This step is crucial for isolating the face from the background and other objects.

Face Analysis: Once the face is detected, an image of it is captured and analyzed. Facial recognition technology typically relies on 2D images because they can be conveniently matched with public photos or images in a database. The software analyzes various facial features such as the distance between the eyes, the shape of the cheekbones, and the contour of the lips and chin. These features help create a unique profile of the face.

Converting the Image to Data: The facial features analyzed are then converted into digital data. This process transforms the analog information of the face into a mathematical formula known as a faceprint. Each person has their own unique faceprint, similar to how fingerprints are unique to each individual.

Finding a Match: The faceprint derived from the captured image is compared against a database of known faces. This database can include images from various sources such as government databases, social media platforms, or private collections. If there is a match found between the faceprint and an image in the database, the system makes a determination of the person's identity.

[10]

Facial Detection Algorithm

An effective facial detection algorithm is crucial for improving facial recognition accuracy. Several methods exist, including geometry-based approaches, feature invariant methods, and machine learning techniques. One of the most effective and widely used algorithms is the Viola-Jones detection algorithm, known for its high detection rate and speed. This algorithm uses an Integral Image and the adaboost learning algorithm as classifiers, and it performs well in various lighting conditions.

IRIS Authentication

Iris recognition captures an image of the iris using a high-resolution camera. The eye section is isolated and segmented, and mathematical algorithms process this image to locate and extract the iris, considering factors like pupil size and eye rotation. Natural variations in iris patterns, such as crypts, furrows, and the collarette, are analyzed to create a unique iris code. Daugman's algorithm, using Gabor filters, extracts texture and converts this code into a binary or numerical representation. This iris template, with its mathematical representation of distinctive features, ensures robustness under various conditions like lighting and pupil dilation, making it a reliable biometric technology​.

Retina Recognition

Retina recognition uses the unique arrangement of cells in the retina for heightened security. Initially conceived by Dr. Carleton Simon and Dr. Isodore Goldstein in 1935, it boasts an impressive recognition rate of about 90%. Recent innovations, such as the method by Sukumaran et al. Using fractal dimension and the system by Tuama and George focusing on retinal vascular diagrams, have demonstrated high accuracy with low computational requirements​.

Hand biometric techniques use the unique features of a person's hand to verify their identity. These methods exploit various physical attributes of the hand, which can be distinctive and difficult to replicate. Here’s a detailed overview of the different hand biometric techniques:

1.     Hand Geometry Recognition

Description: Hand geometry recognition measures the shape and dimensions of a person's hand. It focuses on attributes like the length and width of the fingers, the distance between joints, and the overall shape of the hand.

Process:

Capture: A 2D or 3D scanner captures the hand's shape. This may involve placing the hand on a flat surface or using a 3D camera.

Feature Extraction: Key features such as finger lengths, hand width, and the shape of the palm are extracted.

Matching: These features are compared against a database of enrolled hand shapes to verify identity.

Advantages:

1. Robust and easy to implement.
2. Low susceptibility to changes like aging.

Disadvantages

1. Less distinctive compared to other biometric methods, as hand shapes can be somewhat similar among individuals.

2.     Handprint Recognition

Handprint recognition involves capturing and analyzing the unique patterns of ridges and valleys on the palm, similar to fingerprint recognition but on a larger scale.

Process      

Capture: A high-resolution image of the hand's palm is taken.

Feature Extraction: The system identifies and extracts the unique ridge patterns, minutiae points, and other distinctive features.

Matching: The extracted features are compared with those in a database.

Advantages

1. High accuracy due to the complexity and uniqueness of palm patterns.
2. Effective in identifying individuals even if they are wearing gloves.

Disadvantages

1. Requires high-quality imaging and can be affected by changes in skin condition.

3. Vein Pattern Recognition

Vein pattern recognition analyzes the unique patterns of veins in the hand. This technique leverages the fact that these patterns are unique to each individual and are not easily observable.

Process:

Capture: An infrared camera captures images of the vein patterns beneath the skin.

Feature Extraction: The system extracts and processes the vein patterns.

Matching: The extracted vein patterns are compared to a stored database.

Advantages:

1. High security as vein patterns are difficult to forge.
2. Less susceptible to changes in the outer skin layers.

Disadvantages:

1. Requires specialized imaging technology.
2. Performance can be affected by changes in blood flow or hand positioning.

4. Finger Vein Recognition

Description: Similar to vein pattern recognition but focuses specifically on the veins in the fingers.

Process:

1. Capture: Infrared light is used to capture the pattern of veins in the fingers.
2. Feature Extraction: The vein patterns are extracted and processed.
3. Matching: These patterns are compared with those stored in a database.

Advantages:

1. High accuracy and security.
2. Less intrusive and does not require physical contact.

Disadvantages:

1. Specialized hardware needed.
2. Can be sensitive to factors like finger positioning and skin condition.

5. Palmprint Recognition

Description: Palmprint recognition focuses on the unique features of the palm, including lines, ridges, and minutiae points.

Process:

1. Capture: A high-resolution image or scan of the palm is taken.
2. Feature Extraction: Unique features such as line patterns and ridge structures are extracted.
3. Matching: The features are compared with a database for identity verification.

Advantages:

1. High accuracy and detailed analysis of palm characteristics.
2. Less affected by changes in skin condition compared to some other methods.

Disadvantages:

1. Requires precise imaging technology.
2. Performance can be influenced by hand positioning and movement.

How Finger Vein Recognition Works

Process:

1. Capture:

Infrared Imaging: The system uses near-infrared (NIR) light to capture images of the finger’s vein patterns. The infrared light penetrates the skin and is absorbed by the blood in the veins, making them appear darker than the surrounding tissue.

Sensor Placement: The finger is typically placed on a scanning device where the infrared camera captures a detailed image of the vein patterns.

2.  Feature Extraction:

Image Processing: The captured image is processed to highlight the vein patterns. This involves filtering out noise, enhancing contrast, and extracting the vein structure.

Pattern Analysis: Algorithms analyze the extracted vein patterns, identifying key features such as the arrangement, thickness, and branching of veins.

3. Template Creation:

Feature Extraction: The system creates a biometric template from the vein patterns. This template contains data on the unique features of the vein structure.

Database Storage: The template is stored in a database along with any additional identifiers like names or identification numbers.

4. Matching:

Verification: When an individual attempts to authenticate, their finger vein pattern is captured and processed similarly. The resulting pattern is compared against the stored template in the database.

Decision: The system determines whether the captured pattern matches a stored template to grant or deny access.

Applications of Finger Vein Recognition

Security:

1.     Access Control: Used in secure facilities and high-security areas to control access.

2.     Financial Transactions: Increasingly used in banking and financial services for secure transaction authorization.

Identification:

1.     Government Use: Applied in national ID systems and border control for verifying identities.

Healthcare:

Patient Identification: Used in hospitals to verify patient identities and access medical records.

4. Technical Considerations

Hardware Requirements:

1.     Infrared Camera: Captures the vein patterns using infrared light.

2.     Sensor: Detects the reflected infrared light to generate the vein pattern image.

Software Requirements:

1.     Image Processing Algorithms: Used to enhance and analyze vein patterns.

2.     Matching Algorithms: Compare captured vein patterns with stored templates for authentication.

Performance Factors:

1.     Blood Flow: Variations in blood flow can affect the clarity of the vein patterns.

2.     Environmental Conditions: Temperature and lighting can impact the infrared imaging process.

3.     User Positioning: Accurate placement of the finger is essential for capturing high-quality images.

5. Advantages Over Other Biometrics

1.     Anti-Spoofing: Difficult to forge or spoof, as vein patterns cannot be easily replicated.

2.     Hygiene: Non-contact scanning reduces issues related to contamination or wear.

3.     Stable: Vein patterns remain relatively stable over time, providing consistent performance.

Behavioral Biometric Authentication

Behavioral biometrics verify a user's identity based on unique behavioral patterns like typing rhythm, mouse movements, gait analysis, and voice intonation. This method offers a non-intrusive and user-friendly way to authenticate individuals without relying on physical attributes. However, behavioral biometrics face challenges such as spoofing or imitation attempts​ .

Keystroke Dynamics

Keystroke dynamics analyze how a person types on digital devices to create a distinctive digital signature. Initially proposed by Gaines et al., advancements include the use of fuzzy logic, neural networks, and pattern recognition techniques like the Bayes classifier.

Advantages

1. User-friendly.
2. Fairly unique compared to other methods.
3. Effective for verification.
4. Low cost.

Disadvantages

1. Vulnerable to hacking.
2. Less effective for positive identification compared to verification. [11]

Voice Authentication Technology

Voice recognition technology has evolved significantly over the decades. In the 1990s, the introduction of Hidden Markov Models (hmms) and advancements by companies like Dragon Systems and IBM led to improved accuracy and the ability to handle continuous speech and larger vocabularies. The 2000s saw the rise of Gaussian Mixture Models (gmms) and other statistical techniques, with major strides made by Microsoft and Google in integrating voice recognition into products like Windows Vista and Google Voice Search. The 2010s marked a revolution with deep learning, using Deep Neural Networks (dnns), Convolutional Neural Networks (cnns), and Recurrent Neural Networks (rnns) to significantly enhance accuracy and capabilities. This era saw the development of advanced voice assistants such as Google Assistant, Siri, Alexa, and Cortana. Recent innovations in the 2020s include context-aware and multimodal systems, privacy-preserving techniques like federated learning, and improvements in real-time processing. Today, voice recognition technology is increasingly integrated into iot devices and offers highly personalized, context-aware interactions.

Advantages

1. Hands-Free Operation: Allows users to interact with devices without using their hands.

2. Accessibility: Provides critical support for individuals with disabilities.

3. Efficiency: Speeds up tasks like composing messages and executing commands.

4. Natural Interaction: Feels more intuitive and engaging.

5. Multitasking: Enables users to perform tasks while doing other activities.

6. Streamlined Data Entry: Improves efficiency in fields like healthcare.

Disadvantages

1. Accuracy Issues: May misinterpret accents, noise, or non-standard speech.

2. Privacy Concerns: Risks related to data security and unauthorized access.

3. Context Understanding: May struggle with complex or ambiguous speech.

4. Connectivity Dependency: Requires stable internet for cloud-based processing.

5. Limited Multilingual Support: Quality varies across languages and dialects.

6. Security Vulnerabilities: Susceptible to voice spoofing attacks.

7. Ambient Noise Sensitivity: Performance can be affected by background noise.

Signature authentication

It is a method used to verify an individual's identity and authorize transactions based on their handwritten signature. It involves comparing a signature on a document or transaction with a pre-existing signature on record to ensure authenticity. Whether it is a digital or a physical signature is it widely used in financial transactions, legal documents and to authenticate. Although, it provides security and integrity but there is a risk of forgery and mistakes.

Signature authentication involves capturing a user’s handwritten signature using digital tools, such as tablets or touchscreens, to record both static and dynamic features like shape, speed, and pressure. These features are extracted and encoded into a biometric template, which is then stored in a secure database. During authentication, a new signature sample is captured and compared against the stored template using specialized algorithms to verify identity. The system's accuracy is assessed through metrics like false acceptance and rejection rates, and it is integrated into applications while providing user training and support.

 

                                         Signature [12]

Biometric authentication has become crucial for enhancing security and usability in technology applications. Utilizing unique physiological and behavioral traits, biometric systems provide a reliable way to verify identities, ensuring only authorized users access sensitive information and services. This review covered various biometric technologies, such as fingerprints, facial recognition, iris and retina scans, and voice recognition, highlighting their strengths and weaknesses. Fingerprint identification remains a staple due to its reliability, with advancements in live fingerprint readers enhancing its effectiveness. Facial recognition, supported by algorithms like Viola-Jones, offers a balance of speed and accuracy. Iris and retina scans, known for their precision, are ideal for high-security applications. Behavioural biometrics, such as keystroke dynamics and voice recognition, provide non-intrusive authentication methods, though they face challenges like susceptibility to spoofing and environmental factors.

Integrating biometric authentication into the Internet of Things (IoT) underscores the need for robust encryption and authentication protocols to ensure data integrity and privacy across networks. Its seamless integration into various applications and high security levels make it a valuable tool in the digital security landscape. Future efforts should focus on enhancing system robustness, addressing vulnerabilities, and ensuring widespread user adoption to fully realize the potential of biometric technology.

1. Ambalakat, P. (2005). Security of biometric authentication systems. In 21st Computer Science Seminar (SA1-T1-6), p. 7.
2. Wajgade, V. M., & Gandage, S. C. (2022). A review study of biometric authentication techniques. International Journal of Current Science (IJCSPUB), 12(2), 714.
3. Bhattacharyya, D., Ranjan, R., Alisherov, F. A., & Choi, M. (2009). Biometric authentication: A review. International Journal of u- and e- Service, Science and Technology, 2(3), 13-28.
4. Jaiswal, S., Bhadauria, S. S., & Jadon, R. S. (2011). Biometric: Case study. Journal of Global Research in Computer Science, 2(10), 19.
5. Kanchana, S. (2018). Fingerprint-based biometric authentication in IoT for resolving security challenges. International Journal of Research and Analytical Reviews (IJRAR), 5(4), 1000.
6. Jain, A. K., & Nandakumar, K. (2013). Biometric authentication: System security and user privacy. SIJ Transactions on Computer Science Engineering & its Applications (CSEA), 1(4), 128-135.
7. Kalyani, C. H. (2017). Various biometric authentication techniques: A review. Journal of Biometrics & Biostatistics, 8(5). https://doi.org/10.4172/2155-6180.1000371
8. Kodituwakku, S. R. (2015). Biometric authentication - A review. International Journal of Trend in Research and Development, 2(4), 113-118. Retrieved from
9. Bayometric. "Minutiae-Based Feature Extraction in Fingerprint Recognition Systems." Retrieved from https://www.bayometric.com/minutiae-based-extraction-fingerprint-recognition/.
10. Jubair, Mohammed & Mostafa, Salama & Mustapha, Aida & Hassan, Mustafa. (2019). An Optimized Method for Automated Stadium Attendance Management System. 8. 112-118. 10.14419/ijet.v8i1.7.25965.
11. Maxion, R. (n.d.). Keystroke forensics - Fingerprints in the cyber world. Carnegie Mellon University. Retrieved August 2, 2024, from https://www.cs.cmu.edu/~maxion/projects/KeystrokeForensics.html
12. Offline Signature Verification System Based on the Online Data - Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/figure/Sample-signature-images_fig3_26512856 [accessed 9 Aug 2024]