4 IntrodutionThe ever increasing criminal and terrorist acts in public places, government/private properties resulted in chaotic scenarios thereby making the existing surveillance systems questionable –“How effective is the surveillance system to prevent unlawful activities?”“Does the system guarantee the required level of security?”“Does the system distinguish an unauthorized user from an impostor reliably?”“Is the system user friendly?”.
5 IntrodutionIn early stages, the identity manage- ment systems relied on cryptographic methods requiring the users to remember a secret text (password) or keep something with them (token, card) or a combination of both to prove their identity.Access control, computer logins, checking, making bank transactions, border control, welfare disbursements etc.However, the users challenged the efficacy of such systems.
6 IntrodutionAs an answer, the research community proposed the idea of human identification based on physiological or behavioral attributes of individuals very often termed as “Biometrics”.Although biometrics is not the perfect solution but it offers several advantages over knowledge and possession based approaches in the way thatthere is no need to remember anything,biometric attributes can not be lost, transferred or stolen,offer better security due to fact that these attributes are very difficult to forge and require the presence of genuine user while granting access to particular resources.
7 IntrodutionAny physiological or behavioral attribute can qualify for being a biometric trait unless it satisfies the criteria such asUniversalityDistinctivenessInvarianceCollectabilityPerformanceAcceptabilityCircumvention
9 IntrodutionThe paper provides details about each biometric modality along with prominent recognition techniques and public data sets available to the researchers.The article contains a rough quantitative analysis of individual techniques in terms of accuracy.The paper discusses multi-modal biometrics in detail which has not been well covered in other review articles on the subject.The paper contains current market statistics along with future prospects for biometric based identification and authentication.
11 Biometric systemsA biometric system is a pattern recognition system which matches the salient or discriminatory features of acquired image (probe image) with the features of pre-stored images (gallery image).For doing so, every biometric system comprises ofImage acquisition moduleFeature extraction moduleMatcher moduleDatabase module
12 Biometric systems 2.1. Modes of operation 2.1.1.Verification (positive recognition)The user after submitting the biometric signature to the system claims a parti- cular identity through a PIN, login name, etc.Recognition system validates or voids user's claim by making a 1:1 (one-to-one) comparison between the submitted biometric signa- ture and enrolled biometric signature associated with that parti- cular identity.
13 Biometric systems 2.1. Modes of operation 2.1.2.Identification (negative recognition)Recognition system tries to recognize the user by comparing the submitted biometric signature to all the enrolled signatures in the database by making 1:N (one-to-many) comparisons without specific identity claim from the user.
14 Biometric systems 2.1. Modes of operation ScreeningThis is an extension to identification where the biometric system assures that a particular individual does not belong to a watch list of identities by performing 1:N (one-to-many) comparisons throughout the database.
15 Biometric systems 2.1. Modes of operation ApplicationVerificationComputer logins, ATMs, e-commerce, access control and user authentication on mobile devices .IdentificationIssuance of ID cards, passports, riving licenses, border crossing and welfare disbursements [1,2].ScreeningAirport security, surveillance activities, public place and public events security etc. 
16 Biometric systems 2.2. Performance measurements The performance or accuracy of a biometric system is data dependent usually influenced by environmental and performance factors.Environmental Factors:Temperature, humidity and illumination conditionsPerformance Factors:Capturing good quality images, composition of target user population, time interval between the enrollment and verification phases and robustness of recognition algorithms .
17 Biometric systems 2.2. Performance measurements Sample acquisition errorsEnvironmental conditions surrounding the system give rise to acquisition errors.Failure to enroll errorFailure to capture error
18 Biometric systems 2.2. Performance measurements Performance errorsThese errors are used to measure the accuracy of biometric systems in real environment. A brief description is given below:False match rate (FMR) or false accept rate (FAR)Falsenon – match rate(FNMR)orfalserejectrate(FRR)Equal error rate (EER)
21 Biometric modalities 3.1. Hand region modalities Fingerprints(BEST)PalmprintHand geometryFinger knuckle printFinger nail bedHand vein pattern
22 Biometric modalities 3.1. Hand region modalities
23 Biometric modalities 3.1. Hand region modalities Image acquisition methods:OpticalThermalSiliconUltrasonic imagingExceptions:Hand geometry → use a camera to acquire 3D imageHand vein → beneath the skin thereby exposing hand to near infrared light
25 Biometric modalities 3.2. Facial region modalities FaceEar shapeTeethTongue print
26 Biometric modalities 3.2. Facial region modalities 3D:However, due to the challenges stated above as well as medium level of distinctiveness, facial recognition systems are very much prone to recognition errors.The potential advantages associated with tongue print include protection from external environment and its squirm which proves the liveliness of the subject.The conformity of tongue print as biometric attribute requires large scale studies.牙齒部分主要用在法醫界~ 作者認為超出範圍，故不進一步探討
27 Biometric modalities 3.2. Facial region modalities Image acquisition methods:2D3DFacial thermographlocalization and segmentation
28 Biometric modalities 3.2. Facial region modalities Advantage:The most naturenon-intrusivecontactlessDisadvantage:cosmeticsplastic surgeryperson's face may change or be changed over timeexpensive imaging hardwareillumination conditions, pose variations , ageing conditionspoor recognition accuracyhigher public acceptance
29 Biometric modalities 3.3. Ocular region modalities IrisRetinaSclerapossesses most accurate,highly reliable,well protected,stable and almost impossible to forge biometric signatures,
30 Biometric modalities 3.3. Ocular region modalities
31 Biometric modalities 3.3. Ocular region modalities Image acquisition methods:infra-red light(retina, sclera )illuminating human eye with near infra-red light or in visible wavelength light(iris)position and bifurcations of blood vesselsmeasuring the area of reference (fovea, optic disk)(Sclera)Consequently, the conformity requires large scale studies.IRIS相較非侵入 獲勝!!! retinal scan systems 只用在軍事目的
32 Biometric modalities 3.3. Ocular region modalities Advantage:Highly accurate (higher uniqueness)Highly reliableWell protectedStableAlmost impossible to forgeDisadvantage:synthetic iris images (need iris codes )intrusive (retina)
33 Biometric modalities 3.4. Medico-chemical biometrics Body odor → not realized yetDeoxyribonucleic Acid(DNA)Heart soundElectrocardiogram (ECG)(body odor) albeit some commercial products are already available in the market known as e-noses which can recognize the particular odor such as gases and vapors
34 Biometric modalities 3.4. Medico-chemical biometrics Image acquisition methods:Specialized medical/chemical sensorsSamples from blood, hair, ear wax, dental floss and finger nail clippings(DNA)Different organic compounds(body odor)
35 Biometric modalities 3.4. Medico-chemical biometrics Advantage:The Most accurate (DNA)Disadvantage:IntrusivePrivacy issuesPhysical contact with sensorsRequirement of skilled operatorsFull cooperation from the subjectsDependence over human medical and emotional statesthese of the restrictions towards large scale deployment of such systems.
36 Biometric modalities 3.5. Behavioral biometrics Keystroke DynamicsVoiceSignatureGaitbased on the analysis of the way humans do the things.
37 Biometric modalities 3.5. Behavioral biometrics Voice fixed vocabulary constraints(Gait) Recently, user authenti- cation based on gait analysis has been applied on mobile devices which in turn produced another research arena in gait biometrics.(Gait)用在手機上，手機必須有加速度計這些方法已經開發許久，作者認為其也有不錯的準確度但有人認為 這些作法無法提供充足的獨特性，因為人的習慣及健康狀態與飲食習慣還有老化程度都會影響行為。還有像是背景噪音之類的。肌肉受損也會簽的跟原先不同。此外，步態因為是動態的，也有著很高的誤判綠。無庸置疑的是，人類的行為模式相較之下容易模仿，有些更是十分專業。因此也只能用來做認證用而已
44 Multi-modal biometrics Instead of the tremendous advances in biometric technology, the recognition systems based on the measurement of single modality can not guarantee 100% accuracy. This is due to influence of several factors.Noisy dataIntra-class variationsDistinctivenessNon-universalitySpoof attacksmono-modal VS Multi-modal
45 Multi-modal biometrics 4.1. Sources of information According to the nature of information sources, a multi-modal biometric system can be classified into following categories: [23,275,276]Multi-sensor systemsMulti-algorithm systemsMulti-instance systemsMulti-sample systemsMulti-modal systemsmono-modal VS Multi-modal
46 Multi-modal biometrics 4.2. Modes of operations A multi-modal biometric system functions in two modes:Serial/cascaded modeParallel modemono-modal VS Multi-modal
47 Multi-modal biometrics 4.3. Fusion levels The multiple biometric evidences acquired from different sources can be combined together at different levels:Sensor levelFeature levelMatching score levelDecision levelmono-modal VS Multi-modal
48 Multi-modal biometrics 4.3. Fusion levels mono-modal VS Multi-modal
49 Multi-modal biometrics 4.3. Fusion levels mono-modal VS Multi-modal
50 Multi-modal biometrics 4.3. Fusion levels mono-modal VS Multi-modal
52 Biometrics prospectsThe increasing interest from the government and private stakeholders to adopt biometric based identity management systems prove the efficacy of such systems over the knowledge and possession based systems.
57 ConclusionReliable Identification is a crucial requirement in a wide variety of systems & different scenarios. Traditional cryptographic methods do not offer practical identification capabilities. As an alternative, reliable identification, very often referred as “biometrics” provide better solutions to high security demands of the present day. A wide variety of biometric characteristics have been discovered and tested in recent decades.
58 ConclusionIn this paper, we provided an overview of the existing biometric technologies along with their composition, classification and performance evaluation indicators. This paper also provides an in- depth overview followed by detailed discussion over different biometric attributes along with their potential advantages and challenges. In terms of multiple attributes, careful selection is the key point to success since selecting the modalities belonging to one region may not be a good choice
59 ConclusionFurthermore, we also stress the importance of reliable template protection strategies since the development of artificial gummy fingers and synthetic irises may lead to questions regarding the reliability of the most secure biometric signatures so far.
60 ConclusionDespite the challenges, the market trend presented in the paper portrays a keen interest in biometric solutions to a number of identity management tasks. The large deployment of such systems in private as well as the stringent security environments is evident from the exemplary evidences produced here. Conclusively, the increasing trend of interest from government and private stake- holders promises a bright future to the science of biometrics.