Ear biometrics Advisor WeiYang Lin Professor Group Member

Ear biometrics Advisor: Wei-Yang Lin Professor Group Member: 陳致豪 黃笙慈 695410070 695410128

OUTLINE n n Biometric in general Three kinds of ear biometrics – Burge and Burger – Victor, Chang, Bowyer, Sarkar – Hurley, Nixon and Carter n n Related news Reference

Ideal biometric n Universal : each person should possess the characteristics n Unique : no two persons should share the characteristics n Permanent : the characteristics should not change n Collectable: easily presentable to a sensor and quantifiable

Biometric suitability for authentication purpose [1]

Ideal biometric (cont. ) n Why do we must have ear biometric? – Many problems in face recognition remain largely unsolved. n. A wide variety of imaging problem. n Face is the most changing part of the body. – Facial expression, cosmetics , anaplasty.

Before and after n The magic of cosmetic

Before and after (cont. ) n Anaplasty

Before and after (cont. ) n Anaplasty and cosmetic

Ear shape n Physical biometric is characterized by the shape of the outer ear, lobes and bone structure n Unique enough? n New biometric, not widely used yet n No applications available yet

Alfred Iannarelli n Compared over 10, 000 ears drawn from a randomly selected sample in California n Another study was among identical and non-identical twins – Using Iannarelli’s measurements – Result: ears are not identical. Even identical twins had similar but not identical ears.

Alfred Iannarelli (cont. ) n The structure of the ear does not change radically over time. n The rate of stretching is about five times greater than normal during the period from four months to the age of eight, after which it is constant until around 70 when it again increases. [2]

Permanence of biometrics [1]

Iannarelli’s measurements (a) Anatomy, (b) Measurements. (a) 1 Helix Rim, 2 Lobule, 3 Antihelix, 4 Concha, 5 Tragus, 6 Antitragus, 7 Crus of Helix, 8 Triangular Fossa, 9 Incisure Intertragica. (b) The locations of the anthropometric measurements used in the “Iannarelli System”. (Burge et al. , 1998) [2]

Iannarelli’s system weaknesses n If the first point is incorrect, all measurements are incorrect n Localizing the anatomical points is not very well suitable for machine vision – some other methods had to be found

Methods using pictures (1/3) n Burge and Burger (1998, 2000) – automating ear biometrics with Voronoi diagram of its curve segments. – a novel graph matching based algorithm for authentication, which takes into account the possible error curves, which can be caused by e. g. lightning, shadowing and occlusion. [3]

System step n Acquisition – 300*500 image using CCD camera n Localization – Locate the ear n Edge extraction – Compute large curve segments

System step (cont. ) n Curve extraction – Form large curve segment, remove small ones n Graph model – Build Voronoi diagram and neighborhood graph

Error correct group matching n Compute distance between graph model, if it less than a threshold, identification is verified. n For high FRR due to graph model, we can remove the noise curve and use ear curve width.

Removal of noise curves in the inner ear Graph model (Burge et al. ) and false curves because of e. g. oil and wax of the ear.

Improving the FRR with ear curve widths, an example width of an ear curve corresponding to the upper Helix rim better results

Methods using pictures (2/3) n Victor, Chang, Bowyer, Sarkar (at least 2 publications in 2002 and 2003) – principal component analysis approach – comparison between ears and faces n This method is presented later with 2 cases. [4][5]

Case 1: an evaluation of face and ear biometrics n The used method is principal component analysis (PCA) and the design principle is adopted from the FERET methodology n Null hypothesis: there is no significant performance difference between using the ear or face as a biometric[4]

PCA Method

Points for normalization

Tests of research n For faces: – Same day, different expression – Different day, similar expression – Different day, different expression n For ears: – Same day, opposite ear – Different day, same ear – Different day, opposite ear

Same day, different expression or opposite ear

Different day, similar expression or same ear

Different day, different expression or opposite ear

Victor et al. research result Experiment # Face/Ear compared 1 Same day, different Expected Result Same day, Greater variation in Face performs opposite ear expressions than ears; ears better expression 2 Different day, similar perform better Different day, Greater variation in same ear expression across days; ears better expression 3 Differet day, different expression Face performs perform better Different day, Greater variation in face Face performs opposite ear expression than ear; ears better perform better

Case 2: Ear and Face images n Hypothesis: – ear provide better biometric performance than images of the face – exploring whether a combination of ear and face images may provide better performance than either one individually[5]

Images used in research Same kinds of sets for faces, too. PCA, FERET

Tests for the research n Day variation – other conditions constant n Different lightning condition – taken in the same day in the same session n Pose variation – 22. 5 degree rotation, other conditions constant, taken in the same day

Day variation test

Different lightning conditions

Pose variation (22. 5 degree rotation)

Results n In this research face biometrics seem to be better in constant conditions, ear biometrics in changing conditions n Multimodal biometrics face plus ear gives the best results, why not use them?

Methods using pictures (3/3) n Hurley, Nixon and Carter (2000, 2005) – force field transformations for ear recognition. – the image is treated as an array of Gaussian attractors that act as the source of the force field – according to the researchers this feature extraction technique is robust and reliable and it possesses good noise tolerance.

Error possibilities in ear recognition

Possibilities to enhance ear biometrics n Using accurate measurements, e. g. ear curve and upper helix rim n Removing noise curves n Thermograms removal of obstacles n Better quality cameras more accurate pictures n Combined biometrics

Ear shape applications n currently there are no applications, which use ear identification or authentication n crime investigation is interested in using ear identification n active ear authentication could be possible in different scenarios

Related news n A new type of ear-shape analysis could see ear biometrics surpass face recognition as a way of automatically identifying people, claim the UK researchers developing the system. [6] n University of Leicester working with a Northampton company have made a breakthrough in developing a computerized system for ear image and ear print identification. [7]

Reference n n n n [1] http: //www. bromba. com/faq/biofaqe. htm [2] A. Iannarelli, Ear Identification. Forensic Identification Series. Paramont Publishing Company, Fremont, California, 1989. [3] Biometrics Personal Identification in Networked Society, chapter 13, Mark Burge and Wilhelm Burger [4] Victor, B. , Bowyer, K. , Sarkar, S. An evaluation of face and ear biometrics in Proceedings of International Conference on Pattern Recognition, pp. 429 -432, August 2002. [5] Chang, K. , Bowyer. K. W. , Sarkar, S. , Victor, B. Comparison and Combination of Ear and Face Images in Appearance-Based Biometrics. IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 9, September 2003, pp. 1160 -1165. [6] http: //www. newscientist. com/article. ns? id=dn 7672 [7]http: //www. findbiometrics. com/Pages/feature%20 articles/earprint. html