Egyptian Journal of Forensic Sciences (2016) 6, 99–107

HO ST E D BY Contents lists available at ScienceDirect

Egyptian Journal of Forensic Sciences

journal homepage: http://www.journals.elsevier.com/egyptian-journal-of-forensic-sciences

ORIGINAL ARTICLE

External ear: An analysis of its uniqueness

* Corresponding author. Tel.: +91 9455220839.

E-mail address: r.purkait@gmail.com.

Peer review under responsibility of The International Association of

Law and Forensic Sciences (IALFS).

http://dx.doi.org/10.1016/j.ejfs.2016.03.0022090-536X � 2016 The International Association of Law and Forensic Sciences (IALFS). Production and hosting by Elsevier B.V.This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Ruma Purkait *

Department of Anthropology, University of Allahabad, Allahabad, U.P. 211002, India

Received 19 November 2015; revised 25 February 2016; accepted 21 March 2016Available online 11 April 2016

KEYWORDS

External ear;

Uniqueness;

Euclidean distance;

Multi dimensional feature

space;

Biometrics;

Personal identification

Abstract: In modern times the establishment of personal identity has widened its scope consider-

ably by encompassing various morphological traits e.g. face, iris, retina, hand etc. apart from tra-

ditional dactyloscopy. For any trait to be used for identification purposes it is essential that its

uniqueness is proved in all individuals. The external ear being a new entrant in this field has not

yet been investigated in this aspect.

In the present study an investigation into its uniqueness was undertaken on a large database (1404

adult male and 1257 female subjects) from Central India. Every ear pattern was paired and com-

pared with every other ear present in the sample. Each ear was represented as a feature point in

17 dimensional feature space. The dissimilarity in the ear pattern was measured by Euclidean dis-

tance between members of the pair.

More than 99.9% of ears were found to occupy a distinct position in the multi-dimensional feature

space. Few pairs which could not be distinguished by the above method were successfully differen-

tiated by subjecting their images to direct superimposition. When the left and right ears of the same

individual were compared by the above methods, none of the pairs were left undistinguished. A val-

idation test conducted in North India (on 132 adult males and 168 adult females) supported the

above findings.

Hence, the individuality of every ear has been confirmed which may find use in personal identifica-

tion studies. The study is a step towards providing scientific support for admitting ear evidence in

the Court of Law.� 2016 The International Association of Law and Forensic Sciences (IALFS). Production and hosting by

Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/

licenses/by-nc-nd/4.0/).

1. Introduction

With ubiquitous use of Closed Circuit Television (CCTV),camera phone, digital camera and other electronic gadgets in

the present day life, the comparison of images are more fre-

quently encountered for identification in Forensic cases. Estab-lishing identity from human morphology and behaviour(Biometrics) has been practiced for a long time. With the

advent of computerization and automation there is immenseimprovement in the computation time and thereby identifica-tion process. Apart from the common traits like face, finger-

print, hand geometry, etc. in recent times scientists areexperimenting with new traits to make biometric a wholesomeidentification system. The external ear is a new entrant in thisfield. The possibility of using the external ear as a tool for

100 R. Purkait

establishing personal identity was recognized as early as 1893by Bertillon1 and later the aspect was extensively researchedby Iannarelli2 on a large sample. Since then, several studies

have been reported on ear pattern by computer scientists.3–9

The chances of identification are greatly enhanced if informa-tion regarding the shape of morphological parts of ear, e.g.

shape of ear lobe or any peculiarity like tragal tag, lop ear, Vul-can ear etc. is available. Even the presence or absence of earpiercing itself is a useful attribute in personal identification.10

The ear has few additional advantages over the conventionalbiometric traits, e.g. face. The ear is more coplanar, has uniformdistribution of colour, less affected by ageing, unaffected bychange of facial expression and not affected by facial make-

up like moustaches, spectacle; though hair and ear ornamentscan occlude its appearance.3–7,9 Unlike traits like the iris andretina, imaging of the ear causes less anxiety and can be cap-

tured from a distance.11 The external ear does suffer from afew limitations. It is an established fact that various parts ofexternal ear do exhibit changes with progression of age. The

elongation of the lobule contributes the most to the overallincrease of ear length, especially after 60 years of age.12,13 Thiswill necessitate law enforcement agencies to frequently update

the biometric database for all ages beyond 60 years.Any trait used for establishing personal identitymust possess

few essential properties, one of which is being unique in all indi-viduals. So far little attention has been paid in this direction, and

most of the studies so far undertaken (by computer scientists)3–8

work on the premise that ‘nature never duplicates itself’ or that‘nature creates things and shapes with great variation between

individuals and that there might be enough variation for anear to be individualised’.8 But the Court of Law following stric-ter criteria (Frye v. United States 1923 &Daubert, et al. v. Mer-

rell Dow Pharmaceuticals 1993 rulings)14 demands a morescientific and objective approach for the admission of any evi-dence. This has compelled other age old accepted evidence traits

like fingerprint or documents to conduct uniqueness test in theirfield.14–16 Hence, a preliminary study on a limited sample of 700individuals was undertaken by the author in 2008.17 The studywas based on 12 direct measurements taken on the external

ear. The outcome of the study was quite encouraging. Themajority of themembers of the population studied were well dis-tinguished in the twelve dimensional feature space so developed.

The aim of the present study was to conduct a detailedinvestigation into the uniqueness of the external ear on a largedatabase with increased dimensions of the feature space. Since

images are more frequently encountered in investigations thepresent study was undertaken on them. Improving on the lastattempt17 the external ear was represented by 17 differentparameters measured across its structure. It is a fact that the

larger the number of parameters defining any anatomicalstructure the finer the assessment will be on its morphology.Hence, in an effort to present the morphology of ear in total-

ity, the seventeen distances measured in the study coveredevery possible feature of the ear.

2. Materials and methods

2.1. Subjects

The study presented is a part of the research work undertakenfor the degree of Doctor of Science. The data collection for the

research spanned over a period of four years, 2008 to 2012from all cross sections of people residing in the state of Mad-hya Pradesh (Central India). Three districts out of 51 districts

constituting the state had been covered for the present study.Wards and villages in two tehsils, Sagar and Khurai of Sagardistrict, Tarana tehsil in Ujjain and Begumgunj tehsil in Raisen

districts formed the study area. The majority of the data werecollected by personal house visits; while subjects were alsoapproached at the University campus, colleges, Police acad-

emy, Military cantonment, government offices etc.All the subjects were adults (ranging in age from 20 to

50 years), normal and healthy. None of them suffered fromany auricular (congenital or traumatic) or maxillofacial defor-

mity. Data for the present cross-sectional study were collectedby the non random method. Unrelated subjects belonging toall caste groups and religion in the general population were

inducted. Data were also collected from two tribal groups, Bhiland Saura. Images were procured from 2661 subjects of which1404 belonged to male and 1257 to female subjects. Any image

not conforming to the standard set in the study was rejected.As a validation test the same study was carried out in

Hamirpur district of Himachal Pradesh in North India, a state

located in a different geographical setting (hilly area) with adifferent ethnic composition. 300 adult subjects photographedformed the test sample of which 132 were males and 168females.

The subjects were informed in detail about the study andthose who volunteered to participate were inducted. Participa-tion information sheet explaining the details of the study, its

purpose etc. was prepared in English and Hindi scripts (locallanguage) conforming to the guidelines set by Indian Councilof Medical Research, New Delhi (Fig. 1). The information

sheets were distributed and the details verbally explained tothe subjects. Only those subjects who voluntarily agreed totake part in the study were recruited. Confidentiality and

anonymity was maintained regarding the identity of the sub-ject by allotting a code number (Fig. 2).

2.2. Method

Detailed descriptions of the theoretical background of thetechnique, method of comparison of images and imaging tech-nique adopted have been discussed in an earlier study.18 A

brief description of the experimental method is provided here.The subjects were positioned at a distance of 1.10 m from

the camera with his/her head in the Frankfurt horizontal

(FH) plane. The face of the subject rested on a fabricated ‘Chinstand’. A rectangular scale (60 * 80 mm) was affixed in front ofthe ear so that it was a common tangent to the bulge of tem-poral and zygomatic bone (Fig. 3). Care was taken to keep

the vertical axis of the scale parallel to the vertical plane ofthe ear. The horizontal midline of the scale was positionedto be parallel to the FH plane. During photography the focal

plane of the camera was parallel to the vertical plane of therectangular scale. Bilateral profile images were acquired withKodak Easy Share CX7330, 3.2 Mega Pixel digital camera

using 3� Optical zoom.Using the devised programme ten anatomical landmarks

(superaurale P1, subaurale P2, intertragica inferior P3,

protragion P4, antitragus superior P5, incisura anterior aurisposterior P6, concha superior P7, posterior most point on the

Figure 1 Participant information sheet (A and B).

Uniqueness of external ear 101

antihelical curvature P8, postaurale P9, lobule posterior P10)were identified on each ear image (Fig. 3). Seventeen distances

were computed using the above ten landmarks. With P1 (super-aurale) and P2 (subaurale) as primary landmarks their dis-tances from others (P3–P10) were computed (P1P2, P1P3,

P1P4, P1P5, P1P6, P1P7, P1P8, P1P9, P1P10, P2P3, P2P4, P2P5,P2P6, P2P7, P2P8, P2P9 and P2P10).

In the study each ear pattern was represented as a featurepoint (on the basis of the above distances) in a seventeen

dimensional feature space. If every ear is unique it will occupya distinct location in the feature space and will be separated by

definite Euclidean distance from other ear patterns. Thedevised programme paired every ear with every other ear pre-

sent in the sample and computed the intra-pair distance. Thesmaller the distance (quantitatively) the more the resemblanceof the ear pattern of the respective members.

2.3. Errors in landmark identification

Independent intra and inter-observer tests were performedon ear patterns to measure the experimental error that is

Fig. 1 (continued)

102 R. Purkait

likely to occur while marking the location of the anatomicallandmarks on an image. When an ear pattern is repeated inthe test the computed distance between the feature-points of

the repeated patterns gives a measure of experimental errorlimit and the final results need to be judged on that basis.

In the present study the maximum intra-pair distance var-

ied from 1.9 mm to 3.5 mm and 3.1 mm to 5.1 mm in intra-and inter-observer tests respectively.18 The highest value ofmaximum inter-pair distance was used to set the threshold

limit for the present study. In other words ear patterns forwhich the intra-pair distance was less than 5.1 mm could notbe distinguished by a representation in the present 17-dimensional feature space. These ear patterns were further

compared by direct superimposition technique using Symme-try Perceiving Adaptive Neuronet (SPAN) programme.19

It may further be noted that an ear feature, so defined, is

parity invariant. Hence, it allows a comparison between right

and left ears of the same individual for study on ear bilateralasymmetry.

3. Results

Each ear pattern is represented as a feature vector in a seven-teen dimensional feature space. The Euclidean distance

between such feature points measured their dissimilarity.The first six rows of Table 1 presents the outcome of theuniqueness test performed in Central Indian population sam-

ple. The test was conducted at first for independent samplesof males and females and later on the merged sample repre-senting the total population of the study area. Every ear pat-

tern was paired with every other ear present in the sample andthe distance between members forming the pair was com-puted. The average intra-pair distance varied from 16.6 mm

Figure 2 Participant consent form signed by independent members.

Uniqueness of external ear 103

to 21.4 mm. The details of distribution of the intra pair Eucli-

dean distances for the left ear in males, females and mergedsamples are represented in Fig. 4. In all the samples a peakcan be seen for the distance range of 10–20 mm. It may be

noted that a very small percentage of intra pair distances wereabove 40 mm.

In the independent samples of both sexes, 0.009% of cases

on the left and 0.003% on the right side failed to be distin-guished in 17 dimensional feature space. The percentage ofear pairs unable to cross the threshold level is relatively smallerwhen both samples are merged. Thus, in overall sense, more

than 99.9% of ear pattern pairs could be distinguished in 17dimensional feature space.

The result of the uniqueness test in the validation sample is

presented in the lower half of Table 1 (rows 7–12). The trend ofaverage intra pair distance between ear pairs is similar to theoriginal sample. The percentage of undistinguished pair ranges

from 0.05% to 0.2% and is much higher than that of CentralIndians. This may be due to the small size of the test sample.

It is interesting to note that the average distance between

ear pairs is persistently greater in males as compared to femalesin both samples indicating greater morphological similarity inthe fairer sex. This apparent sexual dimorphism in the distancebetween ear pairs is further supported by the highly significant

result of the t-value when male and female samples are com-pared in both regions of India.

Figure 3 Anatomical landmarks marked on the image using the interactive devised programme. Rectangular scale affixed in front of the

ear is used for photogrammetric rectification of the image.

104 R. Purkait

In the second phase, after comparing ear patterns amongst

unrelated persons, bilateral matching of ears of the same per-son was undertaken (Table 2). A higher number of females(0.39%) than males (0.14%) remained undistinguished in Cen-

tral India while it was reverse in the test sample of North India.In overall performance more than 99% of the cases could bedistinguished in a multi dimensional feature space. The distri-

bution of the bilateral variation in the Central Indian subjectsis represented in Fig. 5. It may be noted that most of the bilat-eral dissimilarity of ears in both sexes are in the Euclidean dis-tance range of five to 20 mm (83% pairs of male and 87% pairs

of female come in this range).It is true that a few ear patterns could not be distinguished

by the 17 dimensional feature vector considered here and it

should not be considered an unexpected event, since the indi-viduality of an ear pattern is not confined to such a measurablefeature vector. In fact, the individuality of an ear pattern is

manifested in the total ear pattern having a complex combina-tion of various segments curved differently that can be simul-taneously compared only by superimposition. Though Hashimet al.20 have discouraged the use of the superimposition tech-

nique for establishing personal identity in real life situations,in our previous study,17 the nearly coplanar ear patterns weresuccessfully dealt with by SPAN and the distinction between

feature-wise undistinguished ear patterns were clearly shown.Hence in the present case the ear patterns that could not be dis-tinguished by the feature vector alone were treated with

SPAN.19 While doing so the geometrically well defined land-marks on the ear patterns were selected as training examplesand after a successful training, SPAN produced superimpos-

able images of the ear patterns. Thus the undistinguished casescould be successfully differentiated. As an example Fig. 6depicts a case of bilateral matching. Left and right ear patternswere separated by a Euclidean distance of 3.39 mm in the sev-

enteen dimensional feature space. Though the resemblance of

the ear patterns was obvious, reconstructing the images in hor-izontal and vertical strip-wise combinations manifested theirdifferences prominently (Fig. 6C and D).

The misalignment in the helix (points 1 to 4), incisura inter-tragica (point 5), anti tragus (point 6) and lobule (point 7) arerepresented in the horizontal strip wise combination (Fig. 6C).

The complete alignment of incisura anterior auris (point 8),hump of tragus (point 9), conchal-antihelical border (point10), anterior lobular border (point 11) and posterior helicalborder (point 12) are striking. The break in the continuation

of the helical (points 1 to 3), anterior conchal-antihelical bor-der (points 4), antitragus (point 5) and posterior lobular bor-ders (points 6 and 7) are well exposed in the vertical

depiction (Fig. 6D). The points 8 to 12 on the vertical strip pre-sent the point of complete alignment.

4. Discussion

Uniqueness or individuality in Forensic Science is very differ-ent from all conventional fields of science. While normal

science looks only between classes, forensic identification pur-posefully looks beyond ‘class characteristics’ and looks withinclasses. That is to say, while normal science is concerned with

establishing regularities, Forensic Science is concerned with thedetection and exploition of irregularities amongst objectswithin classes. Its central assumption is that objects possessenough differences that on adequate inspection one object can-

not be mistaken for another.21

The question of proving uniqueness of forensic trait/evi-dence continues to be a controversial issue. While on one hand

professionals of Law16,22–24 claim that the issue has beenneglected and never rigorously pursued by Forensic Scientists;

Figure 4 Distribution of intra pair Euclidean distance between

left ear pairs of male, female and combined sample in 17

dimensional feature space in Central India.

Table

1Distancesin

feature

space

amongst

allpossible

pairsofearpatternsin

CentralandNorthernIndia.

Populationsample

Sex

Side

Sample

size

Totalpairsanalysed

Average

Euclideanintrapairdistance

(mm)

Number

ofpairsfalling

Stand.Dev.

t-ValueM-F

Maxim

um

Minim

um

Below

threshold

level

Abovethreshold

level

Originalstudy

ML

1404

984,906

19.34

14.17

38.42

*147.54

2.75

84(0.0085%

)984,822(99.99%

)

(CentralIndia)

FL

1257

789,396

16.56

7.06

75.11

2.78

71(0.0089%

)789,325(99.99%

)

MR

1404

984,906

21.42

16.15

31.55

*137.06

3.22

28(0.0028%

)984,878(99.99%

)

FR

1257

789,396

18.50

12.83

133.11

3.29

26(0.0033%

)789,370(99.99%

)

M+

FL

2661

3,539,130

18.96

13.53

147.54

2.75

179(0.0051%

)3,538,951(99.99%

)

R2661

3,539,130

20.84

15.71

139.80

3.29

73(0.0021%

)3,539,057(99.99%

)

Validationstudy

ML

132

8646

21.56

13.54

5.51

*54.30

2.81

19(0.219%

)8627(99.8%

)

(NorthernIndia)

FL

168

14,028

17.03

11.72

46.88

3.28

9(0.064%

)14,019(99.94%

)

MR

132

8646

23.77

14.57

3.91

*59.93

3.71

12(0.139%

)8634(99.86%

)

FR

168

14,028

20.49

15.31

72.04

2.81

13(0.092%

)14,015(99.91%

)

M+

FL

300

44,850

16.55

12.96

75.11

2.78

23(0.051%

)44,827(99.95%

)

R300

44,850

18.50

13.12

133.11

3.29

23(0.051%

)44,827(99.95%

)

*P<

0.001.

Uniqueness of external ear 105

on the other hand Page et al.25 believe that ‘uniqueness isimpossible to prove, and is not anywhere near as relevant assome may claim. There are a few valid reasons to claim

uniqueness, or to continue this fruitless search for whatremains a philosophical ideal.’

Following the recent Court ruling for more rigorous admit-

tance of forensic evidence scientists have undertaken empiricalstudies to prove the uniqueness of various traits. Balding26

claims that it is impossible to prove any human characteristicto be distinct in each individual without checking every indi-

vidual. But undertaking such formidable tasks of testing everyliving person on earth is not feasible. So a large representativesample of a population can work as substitute data22 for the

test as a small step towards investigating uniqueness.Though the study of ear patterns and its variation has been

a matter of observation and study for over a century, till date

there are limited studies with empirical data to establish itsuniqueness. Various studies on ‘Ear Biometrics’ are being car-ried out on the premise that the ear commands as much

authenticity as fingerprints. The present study investigates thispremise so that evidences related to the external ear can begiven enough scientific support to be admitted as credible evi-dence in the Court of Law.

As a test of distinctness the first phase of the investigationwas focused on pairing and matching ear patterns of unrelated(genetically different) subjects. More than 99.99% of cases

could be distinguished by the suggested metric assessment.The remaining 0.01% pairs though were present below thethreshold level were distinguished by direct superimposition

technique (SPAN programme).The next phase compared left and right ears of the same

individual (genetically identical). It is interesting to note thatnone of the ear patterns remained indistinct. 0.14% of male

and 0.39% of female pairs which failed the metric test weresubsequently distinguished by the superimposition technique(Fig. 6).

The method adopted in the present study for testinguniqueness of the external ear requires active expert anthropo-logical intervention for feature detection on the ear pattern.

This is done keeping in view the factors, e.g. the positionand angle of ear placement on the face would vary with eachand every individual, while the computer scientist working at

full automation level generates an algorithm to train the

Table 2 Assessment of bilateral asymmetry of the ear pattern in Central and Northern India.

Population

sample

Sex Sample

size

Total pairs

analysed

Euclidean intra pair distance (mm) No. of pairs falling

Average Maximum Minimum Below threshold level Above threshold level

Original study M 2808 1404 11.87 71.11 3.39 2 (0.14%) 1402 (99.86%)

(Central India) F 2514 1257 11.14 98.12 2.9 5 (0.39%) 1252 (99.6%)

Validation study M 264 132 14.99 130.23 3.23 3 (1.76%) 129 (97.73%)

(Northern India) F 336 168 9.42 43.41 3.74 0 (0%) 168 (100%)

Figure 5 Distribution of intra pair Euclidean distance between

left and right ears of the same person in 17 dimensional feature

space amongst subjects of Central India.

106 R. Purkait

machine to extract questioned anatomical parts and detect fea-

tures. Such an automated decision generated by the machinemay not be acceptable to the Court of Law, unless it is timetested and fool proof.19 This has been particularly stressed

by Iscan27 that for Forensic expert opinions, it is required thatthe final results of the examination should be suitable for aconvincing presentation before the Court of Law.

As stated earlier, a large number of parameters defining an

anatomical feature brings out better assessment of its struc-ture. It may be said that the feature vectors considered in thepresent study would not have provided a complete description

of an ear pattern although every part of the ear had been cov-ered in the study. In fact, a better comparison of an ear patterncould be given by direct superimposition of images of all sub-

Figure 6 Superimposition of left and right ear patterns of the same in

strip-wise superimposition (D) vertical strip-wise superimposition.

jects. The formidable task of superimposition of all cases(3,539,130 matching pairs) was greatly reduced by the feature

vector test which excluded those cases which were grossly dif-ferent from one another.

The study was conducted on a substantially large sample

and validated the results on a test sample belonging to a differ-ent part of the country. None of the ear patterns were found tobe identical in morphology when compared with other individ-

uals, not even its own counterpart. Hence, one can assumethat, external ear pattern could be used as a unique featurefor personal identification.

Though the conclusions drawn are based on the Indian

population sample, the results suggest that a similar outcomemay be expected from other populations as well.

5. Limitations of the study

The ear has nearly a coplanar structure. With vertical or hor-izontal rotation of the head the shape of various ear features

may not be visible in the image for a faithful analysis. So itis necessary to undertake the imaging of the external ear bythe standard method suggested in the study.

For a comparative study it goes without saying that it hasto be conducted on a normal external ear intact with allanatomical features. Genetically deformed ears, e.g. severe

microtia, anotia, dysplastic ear (‘cauliflower’ ear, ‘crumpledear’ without helix or anti helix, Synotia, etc.) should not beincluded in the study.

Funding

None.

dividual. (A) Left ear pattern, (B) right ear pattern, (C) horizontal

Uniqueness of external ear 107

Conflicts of interest

The author has no conflict of interest.

Ethical approval

Informed consent was obtained from all subjects. The study

was approved by the University ethics committee.

Acknowledgement

The author is indebted to Dr. Partho Sinha, Scientist of Foren-sic Science Laboratory, Kolkata, for developing a programme

to analyse the ear pattern in 17 dimensional feature space andpermitting use of SPAN for superimposition of paired images.She is grateful to the subjects for their cooperation during the

study and also gratefully acknowledges the assistance of Dr.Nasir Ahemad and Mr. Sandeep Prasad Daksha in the field.

References

1. Bertillon A. Identification anthropometrique: instructions signale-

tiques. Paris: Masson; 1893.

2. Iannarelli A. Ear identification. Forensic identification

series. Freemont, California: Paramont Publishing Company;

1989.

3. Burge M, Burger W. Ear biometrics. In: Jain A, Bolle R, Pankanti

S, editors. Biometrics: personal identification in a networked

society. Boston: Kluwer Academic Publishers; 1999. p. 273–86.

4. Moreno B, Sanchez A, Velez JF. On the use of outer ear images

for personal identification in security applications. Proc IEEE

Conf Secur Technol 1999;469–76.

5. Victor B, Bowyer KW, Sarkar S. An evaluation of face and ear

biometrics. Proc Int Conf Pattern Recog 2002;429–32.

6. Hurley D, Nixon MS, Carter J. Force field feature extraction for

ear biometrics. Comput Vis Image Understanding 2005;5(3):84–95.

7. Chang K, Bowyer KW, Sarkar S, Victor B. Comparison and

combination of ear and face images in appearance-based biomet-

rics. IEEE Trans Pattern Anal 2003;25:1160–5.

8. Hoogstrate AJ, Heuvel VH, Huyben E. Ear identification based on

surveillance camera’s images. Sci Just 2001;4:1167–72.

9. Sana A, Gupta P, Purkait R. Ear BIOMETRICS: A NEW

APPROACHProc Sixth Int Conf Pattern Recog.

10. Abbas A, Rutty GN. Ear piercing affects earprints: the role of ear

piercing in human identification. J Forensic Sci 2005;50(2):386–92.

11. Amirthalingam G, Radhamani GA. Multimodal approach for

face and ear biometric system. Int J Comput Sci Issues 2013;10

(5):234–41.

12. Purkait R, Singh P. Anthropometry of the normal human auricle:

a study of adult Indian men. Aesthetic Plast Surg 2007;31:372–9.

13. Purkait R. Anthropometry of normal human auricle. In: Preedy

VR, editor. Handbook of anthropometry: physical measures of

human. New York: Springer Science+Business Media; 2012. p.

903–17.

14. Srihari SN, Cha Sung-Hyuk SH, Arora H, Lee S. Individuality of

handwriting. J Forensic Sci 2002;47(4):1–17.

15. Lockhee Martin Company, 50K vs 50K fingerprint Comparison

Test (unpublished study), 1999.

16. Saks MJ, Koehler JJ. The individualization fallacy in forensic

science evidence. Vanderbilt Law Rev 2008;61(1):199–219.

17. Purkait R, Singh P. A test of individuality of human external ear

pattern: its application in the field of personal identification.

Forensic Sci Int 2008;178:112–8.

18. Purkait R. Uniqueness of external ear: a test in genetic relations.

Ann Forensic Res Anal 2015;2(2):1020.

19. Sinha P. A symmetry perceiving adaptive neural network and

facial image recognition. Forensic Sci Int 1998;98:67–89.

20. Hashim N, Hemalatha N, Thangaraj K, Kareem A, Ahmed A,

Fakhuruddin NK, et al. Practical relevance of prescribing super-

imposition for determining a frontal sinus pattern match. Forensic

Sci Int 2015;253, 137.e1–137.e7.

21. Saks M. Merlin and Solomon: lessons from the law’s formative

encounters with forensic identification science. Hastings Law J

1998;49:1069–141.

22. Saks MJ, Koehler JJ. What DNA ’fingerprinting’ can teach the

law about the rest of Forensic science. Cardozo Law Rev

1991;13:361–72.

23. Saks MJ. Forensic identification science: a summary of the theory,

the science, and the legal history. <tapeexpert.com/pdf/forensi-

cidscie.pdf> 2004.

24. Saks MJ. Forensic identification: from a faith-based ‘‘Science” to a

scientific science. Forensic Sci Int 2010;201(1–3):14–7.

25. Page M, Taylor J, Blenkin M. Uniqueness in the forensic

identification sciences—fact or fiction? Forensic Sci Int

2011;206:12–8.

26. Balding DJ. Weight-of-evidence for Forensic DNA profiles. Eng-

land: John Wiley & Sons Ltd; 2005.

27. Iscan MY. Introduction to techniques for photographic compar-

ison: potential and problems. In: Iscan MY, Helmer RP, editors.

Forensic analysis of the skull. New York: Wiley-Liss; 1993.