Biometric application development, such as facial; recognition, has recently become a key priority for smart cities. In addition, researchers worldwide have worked hard on algorithms and approaches that will make these systems useful in everyday Life. Any security measures must ensure the safety of all sen-sitive information. Passwords are the most prevalent form of identification. As information technology and security algorithms advance, however, more and more systems are starting to rely on biometric variables for the recognition purposes (Ouerhani and Alfalou, 2010; Yan and Wang, 2018; Patel and Kale, 2018). These biometric features allow for individual identification based on some aspect of their physiology or behavior. They also offer many benefits. For instance, the necessity for many passwords or secret codes is eliminated if a person is simply ahead of the sensor. Several biometric recognition methods, including iris, fingerprint, voice, and facial recognition, have been implemented recently (Roy et al., 2023).

Measuring and interpreting biological data for authen-tication or identification is the ground of study known as biometrics. Biometrics studies how an individual can be recognized through physical or behavioral traits. Biometric identification systems date back to ancient Egypt. We call "biometrics" the science of identifying individuals through collecting and analy-zing information about their unique physical or behavioral characteristics. From the Ancient Greek (bios) for life and (metrics) for measurement comes our modern English word "biometric," which means "life measurement." The geometry of the face, finger-prints, D.N.A., ears, irises, retinas, and hands are all examples of physical traits. A persons behavior or dynamic measurements might be reflected in their Signature, voice, and gait, all of which are considered the behavioral traits (Jain and Ross, 2016; Jain and Prabhakar, 2004; Jain and Pankanti, 2006). Here is a growing concern for safety in todays world of sop-histicated digital Technology, prompting the creation of various biometric-based personal authentication solutions. Using a persons distinctive behavioral or physical characteristics, biometrics provides a dep-endable and secure identification method. In personal identity systems, fingerprint recognition is the maxi-mum widespread usage of biometrics. Furthermore, fingerprint authentication is one of the safest and most reliable biometric recognition methods. Because fingerprints are permanent and unique to each indi-vidual, they are the best candidate for biometric security systems (Jain and Ross, 2016). A distinctive pattern of interlaced valleys and ridges on the finger surface characterizes a fingerprint. A single curved segment represents a ridge, and a valley is a space between neighboring ridges. The two primary cate-gories automated fingerprint recognition systems fall into are verification and identification. The benefits of biometrics are summarized in Fig. 1; Fig. 2 lays out the benefits and drawbacks of biometrics (Kamlaskar and Abhyankar, 2021).

Fig. 1: Advantages and Disadvantages of Biometrics.

Pattern recognition techniques provide the basis of biometrics. Using biometrics in places like forensics, security, A.T.M.s, smart cards, personal computers, and even networks is becoming increasingly common. When compared to more traditional authentication techniques, biometrics offers more safety. This paper (Sabhanayagam & Senthamaraikannan, 2018) focuses on biometric recognition systems and provides a literature review of other methods of siding. Biometric recognition technologies circumvent the limits of manual recognition systems. Yet, there are constraints on biometrically-based systems that can be addressed as biometric Technology develops. As can be shown in Table, biometrics is used in a wide variety of the contexts.

Fig. 2: Applications of Biometrics.

Automatic airport check-in, access systems, humanita-rian aid operations, and many more can all benefit from iris recognition technology because it is one of the most reliable biometric technologies for human identification and verification. Rings, corona, crypts, contraction furrows, ciliary processes, freckles, and color are the rich textural information (Maltoni & Prabhakar, 2009) that may be extracted from an iris pattern. Iris designs are one-of-a-kind, easily recog-nizable, completely harmless, and remarkably con-sistent over time. The most distinguishing features of a persons iris pattern must be extracted for reliable iris recognition. Thus, it is essential to select an appro-priate feature extraction approach (Sahu and Shrivas, 2013). 

Characteristics of Biometric

Its important to remember that no biometric is perfect. Its not simple to draw parallels. The Researchers characterize the essential characteristic needs of any biometric qualities included in the table as the following: universality, uniqueness (distinctiveness) collectability, permanence, acceptability, performance, and resistance to circumvention (Sabhanayagam and Senthamaraikannan, 2018). These factors are often called the "seven pillars of biometrics" (Sarkar and Singh, 2020). Fig. 3 provides a comparison of these features across popular biometric methods.

Fig. 3: Characteristics of Biometrics.

Individuals fingerprints, irises, faces, hand veins, D.N.A., voices, signatures, gaits, typing patterns, and other biometric modalities can validate or authenticate them with a high grade of certainty. Traditional authe-ntication systems rely on P.I.N.s, passwords, tokens, and the like, all of which can be compromised if hackers access them. Yet, biometric; verification is quickly replacing more traditional forms of authen-tication. Biometrics, which considers a persons unique physical and behavioral; characteristics, offers the similar level of the security as other, more traditional authentication and verification methods (Bolle and Senior, 2013). Individuals can be dependably recog-nized by the biometric verification system using both observable and non-observable characteristics. Biome-tric identification uses physical identifiers, including a persons face, fingerprints, finger shape, iris, vein, retina, voice, and D.N.A., as the depicted in Fig. 4. 

Fig. 4: Examples of Different Biometric Traits.

Five unique components comprise a biometric identity system: a sensor, feature extractor, template database, matcher, and decision maker. As shown in Fig. 2, a typical biometric authentication system may be created in (Wang and Brosseau, 2017).

Biometric; System Modes

Fig. 5 depicts the two modes of process that the bio-metric authentication system can operate in enroll-ment and verification. There are two sub-steps in the authentication procedure: verification and identify-cation. In addition, positive; and negative identification are distinguished. These modes will be discussed at length in the following section (Jung and Heo, 2018). 

Fig. 5: Modes of Biometric system.

Using Biometric Methods

Biometrics is used to identify and verify individuals by measuring and comparing specific physical traits. Biometrics encompasses many methods that cannister be roughly broken down into two classes. Behavior and physiological traits are depicted in the Fig. 6. This study discusses fingerprint, iris, and facial biometrics on the physiological side and voice, Signature, and keystroke biometrics on the behavioral side (Jung and Heo, 2018).

Fig. 6: Types of Biometric Techniques.

Types of Biometric; Identifiers

Fingerprint Recognition 

One of the greatest widespread and long-standing identification methods today is fingerprinting. Finger-prints are unique to each person because they feature a complex pattern of lines, arches, loops, and whorls. To implement this method, an ink or digital scan of a persons fingertips records information about their fingerprints. Records of fingerprints are processed or saved as an image so that the minutiae, including whorls, arches, and loops, can be compared with other fingerprint data. This method involves the user gently pressing his finger in contradiction of a small reader surface (optical or silicon) for less than 5 seconds during the verification process. The reader is only around 2 inches square. A computer called a reader receives information from the scanner, which subse-quently transmits the data to a database for compare-son. The US, along with other countries, including Canada and the United Kingdom, use a record of fingerprint procedures called Automatic Fingerprint Identification System (AFIS). Fingerprints are an individuals Signature. Despite its importance, dry skin, a poor environment, or an injury might render this procedure ineffective despite its great depen-dability, precision, and uniqueness. There is now a big database of comparative specimens, and the search for matching can be finalized quickly thanks to modern fingerprint procedures backed by computer and laser technologies.

Face Recognition 

A digital video camera is used in this technique to examine the details of still photos of a persons face. It considers every aspect of a persons face, from the distances between their eyes and nose to the widths of their mouths and jaws. When an operator poses in front of the camera, these readings are compared to those already kept in the database. The achievement of this method has limited its usage to verification systems. The participant stands about two feet from the camera, with their back to it. When a being logs in, the scheme looks for their face and checks it against their claimed identity or a facial database. The Facial Recognition Technology Database (FERET)s primary goal is to advance autonomous facial recognition capabilities that cannister be used for security purposes. Verifi-cation takes less than 5 seconds. However, the user may need to adjust his face slightly before it works. Although widely adopted, this inexpensive Technology can be deceived by identical twins and age-related facial changes.

Essential Elements of a Facial Recognition System

Fig. 7 the three main phases of engineerings auto-mated face recognition problem, from left to right: (1) approximate face detection and normalization; (2) feature extraction and correct face normalization; and (3) classification (verification or identification) (Chi-haoui and Ben Amar, 2016).

Fig. 7: The standard design of an automatic face-recognition system.

Iris recognition

Scannable features in the iris diaphragm, the pigm-ented tissue about the pupil, include over 200 data points. The user aligns him such that his reflected image appears on the screen. The iris scanner desires to be within 12-18 inches of the user, unlike the retinal scanner, which needs to be much closer. Since the user needs glance into the gadget, the Verification time is typically under 5 seconds. On the additional hand, this represents the users actual iris pattern, either on a physical I.I.D. card or in a centralized database. The iris region of human eyes is captured in this photo database by a visible-light-operating sensor. If here is a correlation, the user is verified as legitimate. Iris scan-ning and recognition are quick and painless. When likened to fingerprints, it might be more FAR-effec-tive. Iris recognition is more individual than finger-printing but less so than retinal scanning. With almost 240 reference points for a match, this Technology is far superior to the fingerprint method, which only employs 60 points. Unlike fingerprinting, which requires physical contact, this method does not. Theyre better and safer than the alternatives, but theyre also quite costly and require a high memory Fig. 8 shows iris recognition Schemes.

Voice Recognition

Pitch, tone, frequency, and other nuances of the humanoid voice are analyzed and used by speech recognition software.

Fig. 8: Sample iris recognition Systems.

Different vocal tract shapes and acquired speech pat-terns are the primary emphases of this method. To use this method, the user speaks a command into a micro-phone connected to the device. Over twenty factors, including pitch, speech, energy density, wave-forms, etc., are used to investigate their voice and extract useful information. This real-time profile is compared to a previously recorded one in a central database. If here is a good match, the user is verified as legitimate. Although voice recognition is one of the easiest ways because it requires no training and costs very little to implement, it can be problematic in less-than-ideal conditions (such as when its too cold or hot outside). When a persons voice changes, its partly due to their appearance and habits. Male and female vocal cords vibrate at a rate of 80 and 400 times per second, respectively. Some characteristics that donate to the uniqueness of each persons voice are the scope of their jaw opening, the shape and position of their tongue, and their lips.

Signature Recognition

Signature recognition was the least reliable biometric method. In a signature, the text used is continuous and uniform. A tablet or paper is placed over a sensor tablet, and the user signs there. The verification process takes roughly 5 seconds while the device stores the users Signature and comparisons it to its database. The technologies are promoted through a cheap writing tablet, which greatly increases the biometrics cost-effectiveness without considerably reducing its performance accuracy. This Technology has minimal uniqueness despite being inexpensive, non-intrusive, widely accepted by users, requiring little training, and evolving.

Functioning of a Biometric System

This systems two modes of operation training and classification are both straightforward. Todays active-ties are commonly observed, such as the operation of a fingerprint sensor on any cellular telephone. The saved fingerprint pattern is verified every time a finger touches the sensor. The diagrammatic depiction of this mechanism in action is provided below in Fig. 9. In this context, "enrollment" describes the primary step. The data is saved permanently within the system. After implementation, all of the information in the system is double - checked against what was gathered during enrollment to ensure accuracy.

Fig. 9: Functioning of a Biometric System.

We need to make sure that the data is entered and stored safely. In the initial Building Block, the sensor acts as a user interface, gathering information from the outside environment. Its mostly a representation, but that might shift dependent on the systems use. The data from the sensor is modified and de-barred from further processing in the second Block. Important and necessary features are extracted in the following Feature Extractor Block. At this stage, retrieving the right features as efficiently as possible is crucial. The following Block, Template generator, uses the pictured integers with specific attributes to generate templates. All relevant structures from the source have been rolled into this single template. Biometric measure-ments are segregated in templates to reduce file size and maintain individuality. Now that the template has been created, it may be protected in a database, matched with others, and sent to a human who will compare it to the currently active template and deter-mine the degree of deviation using the algorithm. The generated data is output by the matching program after it has accessed the template (Sachdeva, 2021).

Advantages and Disadvantages of the Biometric Techniques

Many modern security systems make usage of Bio-metry identification methods. There are benefits and drawbacks to all of these methods. Based on the nature of the task at hand, the appropriate application and method can be chosen from Table 1.

Table 1: Advantages, Disadvantages and Applications of Biometrics methods.

Review of Literature

In this part, the author discusses various previously published research papers on biometric systems:

In the present work (Sudar and Nagaraj, 2019). We have evaluated the utility of biometric procedures to more conventional authentication methods. Finger-print, iris, retina, face, palm, voice, Signature, and gait biometrics are fair a few of the broad biometric methods weve covered, along with their benefits and downsides. General biometric approaches for security systems are compared and analyzed in this study. This study (Sabhanayagam and Senthamaraikannan, 2018) summarizes the various biometric approaches, discus-sing their benefits and drawbacks. The fingerprint is a fast and accurate biometric approach for a more dependable and secure system, according to a literature review that noted the differences and implications of error rates across other biometric techniques. The natural properties of the iris biometric modality lend credibility and non-intrusiveness to the iris identi-fication system. Current efforts in this field pose difficulties for both small template sizes and quick verification techniques. The act of the iris recognition system has been optimized by making special efforts to decrease the size of the retrieved characteristics. We present a feature fusion method based on multilinear subspace learning to better examine Iris recognition. This strategy has four distinct phases. In the first stage, the iris is segmented out of the eyes visual image. As wavelet packet decomposition may provide good time and frequency resolutions concurrently, it is used in another stage to excerpt features of the iris image. After the nodes or packets have been separated, they are organized into a third-order tensor rather than a lengthy vector, and multilinear principal component analysis (MPCA) is then used to implement feature fusion (Kamlaskar and Abhyankar, 2021) directly. Extraction of tensor features using MPCA has found widespread use in many different computer image and design recognition tasks. Recognizing faces (Vasilescu and Terzopoulos, 2003), processing signals (Plataniotis and Venetsanopoulos, 2011), reading handwriting (Wang and Wang, 2016), recognizing digital numbers (Haiping and Plataniotis, 2008) analyzing content (Muhsain, 2011) and finding outliers in data (Jing, 2019) are all relatively new uses. The authors offer a new MPCA framework for dimensionality reduction and feature extraction of the tensor object and use it as an example in gait identification. Inspired by MPCAs effectiveness in feature extraction, we propose tensor-based MPCA feature fusion for Iris recognition. While the input samples of the back-propagation approach are coded and normalized using an algorithm working using a predefined codebook, N.F. (Muhsain, 2011) explored the challenge of minimizing the fingerprint features given into the neural network. The main benefits of making a codebook are its ease of use and quick processing time. Pre-processing for image enhancement, binarization and thinning of fingerprint images, feature extraction from the thinned image ridge, and finally a matching stage in which similarity and distance measurements are used to match two-minute points are all part of (Jing, 2019) s proposed multi-stage fingerprint recognition system. In (Bakheet and Youssef, 2022) the use of ridge termination and ridge bifurcation as minutiae in fingerprint recognition using artificial neural networks (ANNs) is proposed. Safely extracting the details from the binary fingerprint images is the most important stage in automatic finger-print matching. The presented method improves recognition rates while reducing mistake rates. The experimental findings demonstrated a recognition rate of 91.10 percent on the average using the planned technique. In (Bakheet and Youssef, 2022) ridge ter-mination and ridge bifurcation are used as distin-guishing features in an ANN-based fingerprint recog-nition algorithm. The most crucial part of automatic fingerprint matching is obtaining the details from the binary fingerprint images in a secure manner. The presented method improves recognition rates while reducing mistake rates. (Gronkiewicz and Mickiewicz, 2016) work is among the most reliable approaches for iris recognition. Iris structures are extracted using quadrature 2D Gabor wavelets, and the resultant image is encoded into a 256-byte (2048-bit) binary code. The degree to which two iris codes are similar is measured using the Hamming distance. A Laplacian of Gaussian filter is used in the Iris recognition system proposed in (Wildes, 2021). Finally, I built a Laplacian pyramid to make a small iris pattern. A way to describe an iris recognition system and explaining how to classify data employing a specific strategy was proposed by (Kaur, 2014). Iris recognition on demand is utilized in the academic world to ensure authenticity. The iris recog-nition systems inner workings must be spelled out in this study, but the corresponding approaches still have a way to go before theyre truly robust. Many bio-metric functions were described by using a single parameter, as shown by the work of (Kaur and Verma, 2014). The first step in using a biometric function is extracting the consistent feature. A separate can be identified using the bio-metric systems parameters. Design recognition is the basis of biometrics. The military, forensics, controls, access, and other fields all use on-demand Biometric Technology. Dependent on the context, iris recognition can be an effective bio-metrics method. In (Saini and Rana, 2014) compare various biometric system methods. This article has to be defined as a quick introduction to various metho-dologies, which has to be utilized in an earlier paper and define the comparison by testing the performance on various types of databases and then classifying by various parameters. It has also specified the recogni-tion techniques. Although biometrics security systems have various concerns like data privacy, bodily pri-vacy, spiritual conflicts etc., they nonetheless have benefits that may improve our life magnificently by boosting security.

Simulation Resulting Table

Table 2: Comparative Analysis of Various Biometric Techniques.

Several biometric methods are compared above for their precision, false acceptance and rejection rates, uniqueness, performance, acceptability, security, etc.

In general, biometrics is based on pattern recognition techniques; moreover, it is an emerging technology widely used in security, forensics, smart cards, net-works, personal computers, and A.T.M.s. Biometric verification is more secure than conventional authen-tication methods, besides several biometric approaches are analyzed and contrasted in this research, such as security, performance, precision, uniqueness, accept-ability cost, acceptance rate, rejection rate, and many more which are already mentioned in above table widely. In light of the methodologies above and the comparison table, the author has concluded that the fingerprint technique is the fastest and most accurate biometric technique for more dependable and secure system based on performance and fast communication. Whereas, speech, face, and signature biometric tech-niques were mostly accepted by users. Due to the unique characteristics of the iris (the iris approach delivers the most secure performance, accuracy, uni-queness, and acceptability of all biometric procedures). It can be also used forever as a password. Finally, Iris is the only part of a human that cannot be changed and provides the finest answer overall.

The author(s) would like to thanks the participants and the members of researchers families who provided support and assistance.

 The author(s) confirm that this study has no potential conflict.