207
3C Tecnología. Glosas de innovación aplicadas a la pyme. ISSN: 2254 – 4143 Edición Especial Special Issue Marzo 2020
HUMAN 2D EAR BIOMETRIC RECOGNITION BASED ON
CONTOUR MATCHING TECHNIQUE
Alagarsamy Santham Bharathy
Department of Electronics and Communication Engineering,
School of Electronics and Electrical Technology,
Kalasalingam Academy of Research and Education,
Krishnankoil, Virudhunagar Dt., (India).
E-mail: santhembharathy@gmail.com ORCID: https://orcid.org/0000-0003-0978-3905
Kalpana Murugam
Department of Electronics and Communication Engineering,
School of Electronics and Electrical Technology,
Kalasalingam Academy of Research and Education,
Krishnankoil, Virudhunagar Dt., (India).
E-mail: drmkalpanaece@gmail.com ORCID: https://orcid.org/0000-0002-5121-0468
Recepción:
05/12/2019
Aceptación:
03/01/2020
Publicación:
23/03/2020
Citación sugerida:
Bharathy, A. S., y Murugam, K. (2020). Human 2D Ear Biometric Recognition Based on Contour
Matching Technique. 3C Tecnología. Glosas de innovación aplicadas a la pyme. Edición Especial, Marzo 2020,
207-217. http://doi.org/10.17993/3ctecno.2020.specialissue4.207-217
Suggested citation:
Bharathy, A. S., & Murugam, K. (2020). Human 2D Ear Biometric Recognition Based on Contour
Matching Technique. 3C Tecnología. Glosas de innovación aplicadas a la pyme. Edición Especial, Marzo 2020,
207-217. http://doi.org/10.17993/3ctecno.2020.specialissue4.207-217
208 http://doi.org/10.17993/3ctecno.2020.specialissue4.207-217
3C Tecnología. Glosas de innovación aplicadas a la pyme. ISSN: 2254 – 4143 Edición Especial Special Issue Marzo 2020
ABSTRACT
This paper presents, the Ear detection biometric is obtainable utilizing normal ear method
to detection, which is motivated through normal face acknowledgment methods. This work
proposed another ear correlation method dependent on template expansion. The work is
connected with ear database given by USTB China on which, the work delivered 100%
exactness more than 180 ear images.
KEYWORDS
Image Processing, Ear Images, Feature Extraction, Contour Matching.
209 http://doi.org/10.17993/3ctecno.2020.specialissue4.207-217
3C Tecnología. Glosas de innovación aplicadas a la pyme. ISSN: 2254 – 4143 Edición Especial Special Issue Marzo 2020
1. INTRODUCTION
In the recent science technology biometrics where an element is recognized based on
physical highlights or conduct qualities (Basit, Javed, & Anjum, 2005). Physical attributes
incorporate unique face, retina, nger, palm print, iris, and ear with so forth while
behavioral qualities comprise of step acknowledgment, voice, odour acknowledgment,
with mark conrmation. The acquired biometric outcomes are utilizing solo or dierent
methods. The accomplished outcomes show that biometric methods be considerably extra
exact with precise over conventional systems. But accuracy, it has been dependably sure
issues which stay related to current customary methods. For instance, think about belonging
and information. Both can be shared, stolen, overlooked, copied, lost or removed. Anyway,
the peril is limited in the event of biometric implies (Moreno, Sanchez, & Velez, 1999).
The biometrics work is amiable within a wide range of safety frameworks. By means of the
dangers/progresses of innovations, and it’s needed a constant to look at new methods for
utilizing like remain solitary relevancies or related to current frameworks. To incorporate any
new category of biometric, the state necessary is that it ought to be general, unmistakable,
eternal and collectible for example every people should have those highlights (widespread)
and highlights ought to recognizable in support of every person (particular). The highlights
ought didn’t to shift (everlasting) and it must be anything but dicult to get required
data from these highlights (collectible) (Jain, Hong, & Pankati, 2000). Clearly, ears are an
unmistakable element of all people making it all around satisfactory. Ear biometrics has a
few points of interest over whole face: decreased position able goals, a progressively equal
appropriation hue and reduce uctuation with demeanors and direction of face. In this
proposed work, another ear acknowledgment strategy is planned dependent by and large
ear; it is connected for individual ID. The remaining of this paper is sorted out as pursues.
In section 2 foundation and related work regarding ear acknowledgment are given. Section
3 incorporates pre-handling pursued by highlight origin and coordinating in section 4. The
section 5 test outcomes with talk are accounted for an indenite section 6 ends be made.
2. RELATED WORK
The rst ear was utilized for acknowledgments for individual was elaborated in Iannarelli
(1989) who utilized labor-intensive methods toward distinguish ear pictures. Tests of
210 http://doi.org/10.17993/3ctecno.2020.specialissue4.207-217
3C Tecnología. Glosas de innovación aplicadas a la pyme. ISSN: 2254 – 4143 Edición Especial Special Issue Marzo 2020
more than ten thousands ears were concentrated to demonstrate the uniqueness of ears.
Arrangement of ear could not modify profoundly after some time. The restorative writing
(Victor, Bowyer, & Sarkar, 2002) gives data that ear development is corresponding later
than initial 4 months in birth and modies are not detectable from the age eight to seventy.
In paper, Chang, Bowyer, Sarkar, and Victor (2003), and Chen and Bhanu (2005) utilized
Eigen ear image for distinguishing proof. The outcomes got be diverse in the two types. In
Kumar (2012), Miyazawa, Ito, Aoki, Kobayashi, and Nakajima (2008), Ito, Iitsuka, and
Aoki (2009), Ansari and Gupta (2007), and Hurley, Nixon, and Carter (2005) outcomes
demonstrate no distinction in face and ear execution as Victor’s outcomes demonstrate that
ear execution is more awful over face. As per in Yan and Bowyer (2007), Joshi and Chauhan
(2011), Gonzalez, Woods, and Eddins (2004), and Tang (2016), the distinction in result may
be because of utilization of various picture quality. As in Kumar (2012), utilized 2D force
pictures of ears by means of 3 neural methodologies (Weighted Bayesian, Bayesian, Borda)
for acknowledgment. In this work, three pictures of every individual as of 60 individuals
were utilized to assess the acknowledgment.
3. PROPOSED SYSTEM
Capture Ear Image
Normalize
Feature Extraction
Training
Matching
Decision
Figure 1. Process.
211 http://doi.org/10.17993/3ctecno.2020.specialissue4.207-217
3C Tecnología. Glosas de innovación aplicadas a la pyme. ISSN: 2254 – 4143 Edición Especial Special Issue Marzo 2020
Each picture is experienced the accompanying strides before highlight extraction. Ear
picture is edited physically from the caught head picture of an individual. Edited ear picture
is resized. The hued picture is changed over into grayscale. Concentrate the highlights
structure grayscale picture utilizing vigilant edge nder and spare as the paired picture.
Manual trimming has been done in the work in light of the fact that robotized ear editing
is under the procedure. The sizes of the edited ear picture are extraordinary. So as to
locate a similar amount of highlights as of every ear picture, rearranging the pictures to a
remarkable xed size of 80*150 pixels is completed. Every picture is changed over as of
color to grayscale. At that point, it is sent in support of the component origin part by the
Canny edge nder. In Figure 2 exhibits yield toward nish the pre-handling step. In Figure
2(a) demonstrates the genuine picture in the catalog with the trimmed picture is obvious
in Figure 2 (b). Figure 2 (c) and Figure 2 (d) are the resized edited picture with color and
grayscale individually. Figure 2 (e) is the genuine element removed after pre-preparing.
Figure 2 (a). Real Image.
Figure 2 (b). Image Cropped. Figure 2 (c). Image Resized.
212 http://doi.org/10.17993/3ctecno.2020.specialissue4.207-217
3C Tecnología. Glosas de innovación aplicadas a la pyme. ISSN: 2254 – 4143 Edición Especial Special Issue Marzo 2020
Figure 2 (d). Image Gray Scale. Figure 2 (e). Image Binary with real Features.
3.1. FEATURE LEVEL EXTRACTION WITH MATCHING
While the portioned ear can be legitimately utilized during the coordinating stage, most
frameworks separate a striking arrangement of highlights to speak to the ear. Highlight
extraction alludes to the procedure where the sectioned ear is diminished to a numerical
model (e.g., an element vector) that abridges the prejudicial data. In the wake of normalizing
the ear pictures, the following stage is including extraction and coordinating. Existing
technique vigilant edge nder has been utilized for highlight extraction. Another strategy
is proposed for ear correlation dependent by and large ear picture. In this methodology,
every ear picture highlights are as a parallel lattice of 80×150. To build the thickness of ear
picture highlights, enlargement activity has been achieved on each ear picture. In Figure
3(a) is the real ear picture highlights and 3(b) is the enlarged picture. ‘N’ is a number of
the expanded twofold picture of a similar individual with various variety has been utilized
to gure normal ear picture. Determined normal ear picture has been spared as a twofold
framework for the layout. These exploration works, 180 ear pictures of 60 people (three
pictures of every individual) has been utilized. The 3 pictures of every individual have been
utilized for normal picture guring and spare as a double framework of 80×150 which is
utilized as a layout. The ensuing calculation has planned for ear perceiving.
Stage 1: calculate complete number of pixel in twofold normal ear picture format.
Stage 2: achieve bitwise intelligent OR activity among the normal double picture and
inquiry picture. Tally yet again the number of resultant.
Stage 3: the all-out number of ones include in Stage 2 is same, which is included in stage1,
at that point show the note ear is perceiving through the personality of layout and outlet.
213 http://doi.org/10.17993/3ctecno.2020.specialissue4.207-217
3C Tecnología. Glosas de innovación aplicadas a la pyme. ISSN: 2254 – 4143 Edición Especial Special Issue Marzo 2020
Stage 4: if all out no of one’s includes in Stage 2 is fewer, at that point and equivalent to the
quantity of include in stage1 in addition to limit esteem (for this situation edge worth is 200
pixels) at that point question ear picture is perceived and exit.
Stage 5: Check on the o chance that it is last normal ear layout, on the o chance that
indeed, at that point go to Stage 6 generally go to Step 1 and contrast question picture and
another ear format.
Figure 3 (a). Image in Binary by means of real Features. Figure 3 (b). Image Dilated Features.
Figure 3 (c). 3 Dilated picture utilized in Average Image Calculation. Figure 3 (d). Real Image Average.
4. RESULTS AND DISCUSSION
The planned strategy is actualized in MATLAB 2017 version on a laptop. In the
examination, ear database from the USTB has been utilized. The databases enclose a sum
of 200 pictures with 80×150 pixels goals. A lot of 60 individuals has been utilized for
examinations having at least three pictures each. Three pictures of every individual have
been utilized for normal picture computation. The resultant picture has been utilized as a
format for ear acknowledgment.
214 http://doi.org/10.17993/3ctecno.2020.specialissue4.207-217
3C Tecnología. Glosas de innovación aplicadas a la pyme. ISSN: 2254 – 4143 Edición Especial Special Issue Marzo 2020
Figure 4. Accuracy comparison Ear Recognition.
Time investigation has demonstrated Figure 4, the time examination among time in second
with time taking to perceive explicit ear, at 2.27 GHz Intel 3i processor, inquiry ear picture
correlation time and putting away normal ear format is 0.108 Second. In this analysis, 180
ear pictures of 60 people, three pictures of every individual have been utilized for normal
picture count. The ear acknowledgment rate is 100% percent more than 180 pictures. 20
pictures of 20 people, which isn’t taking an interest in normal picture computation likewise
delivered 90% exactness by utilizing a limit esteems TH= 173. In this examination work,
test on possess database is under-preparing, It is normal that as the number of ear picture
increment for normal picture computation, the acknowledgment rate will increment.
5. CONCLUSION
Ear biometrics got consideration regarding the examination as of late. In this paper, another
technique for human acknowledgment is proposed dependent by and large ear pictures.
Ear pictures are trimmed physically and resized to a xed size pursued by the change
to grayscale. After that Canny edge identier is utilized to remove the element from the
picture. Database pictures are prepared and put away as a normal ear picture. Results got
are promising and empowering with right acknowledgment rate just as the time required.
Results will get better if number of ear pictures increment in normal picture count.
215 http://doi.org/10.17993/3ctecno.2020.specialissue4.207-217
3C Tecnología. Glosas de innovación aplicadas a la pyme. ISSN: 2254 – 4143 Edición Especial Special Issue Marzo 2020
REFERENCES
Ansari, S., & Gupta, P. (2007). Localization of ear using outer helix curve of the ear.
International Conference on Computing Theory and Applications (ICCTA), pp. 688–692.
https://doi.org/10.1109/ICCTA.2007.82
Basit, A., Javed, M. Y., & Anjum, M. A. (2005). Ecient iris recognition method for
human identication. International Scholarly and Scientic Research & Innovation, 1(4), 24-
26. https://publications.waset.org/15966/pdf
Chang, K., Bowyer, K. W., Sarkar, S., & Victor, B. (2003). Comparison and
Combination of Ear and Face Images in Appearance-Based Biometrics. IEEE
Transactions on Pattern Analysis and Machine Intelligence, 25(9), 1160-1165. https://doi.
org/10.1109/TPAMI.2003.1227990
Chen, H., & Bhanu, B. (2005). Contour Matching for 3D Ear Recognition. Seventh IEEE
Workshops on Applications of Computer Vision (WACV/MOTION’05), 1, 123-128. https://
doi.org/10.1109/acvmot.2005.38
Gonzalez, C. R., Woods, R. E., & Eddins, S. L. (2004). Digital Image Processing Using
MATLAB (R). Pearson Education.
Hurley, D. J., Nixon, M. S., & Carter, J. N. (2005). Force eld feature extraction for
ear biometrics. Computer Vision and Image Understanding, 98(3), 491–512. https://doi.
org/10.1016/j.cviu.2004.11.001
Iannarelli, A. V. (1989). Ear Identication. Paramont Publishing Company.
Ito, K., Iitsuka, S., & Aoki, T. (2009). A palmprint recognition algorithm using phase-
based correspondence matching. In 16th IEEE International Conference on Image Processing
(ICIP). https://doi.org/10.1109/ICIP.2009.5413740
Jain, A., Hong, L., & Pankati, S. (2000). Biometric Identication. Communications of the
ACM, 43(2), 91-98. https://doi.org/10.1145/328236.328110
216 http://doi.org/10.17993/3ctecno.2020.specialissue4.207-217
3C Tecnología. Glosas de innovación aplicadas a la pyme. ISSN: 2254 – 4143 Edición Especial Special Issue Marzo 2020
Joshi, K., & Chauhan, N. C. (2011). A Template Matching and Support Vector
Machine based Approach for Human Ear Detection and Varication. International
Conference on Information, Signal and Communication. https://www.researchgate.net/
publication/281741782_A_Template_Matching_and_Support_Vector_Machine_
Based_Approach_for_Human_Ear_Detection_and_Verication
Kumar, A. (2012). Can We Use Minor Finger Knuckle Images to Identify Humans? IEEE
Fifth International Conference on Biometrics: Theory, Applications and Systems (BTAS), pp. 55-
60. https://doi.org/10.1109/cvprw.2014.21
Miyazawa, K., Ito, K., Aoki, T., Kobayashi, K., & Nakajima, H. (2008). An Eective
Approach for Iris Recognition Using Phase-Based Image Matching. IEEE transactions
on pattern analysis and machine intelligence, 30(10), 1741-1756. https://doi.org/10.1109/
TPAMI.2007.70833
Moreno, B., Sanchez, A., & Velez, J. F. (1999). On the Use of Outer Ear Images for
Personal Identication in Security Applications. Proceedings IEEE 33rd Annual 1999
International Carnahan Conference on Security Technology (Cat. No.99CH36303). https://doi.
org/10.1109/CCST.1999.797956
Tang, T. (2016). Biometric Recognition Using Wavelet Domain Features. Electrical and
Electronics Engineering An International Journal (ELELJU), 5.
Victor, B., Bowyer, K., & Sarkar, S. (2002). An Evaluation of Face and Ear Biometrics.
Proceedings 16th International Conference Pattern Recognition, pp. 429-432. https://doi.
org/10.1109/ICPR.2002.1044746
Yan, P., & Bowyer, K. W. (2007). Biometric recognition using 3D ear shape. IEEE
Transactions on Pattern Analysis and Machine Intelligence, 29(8), 1297-1308. https://doi.
org/10.1109/TPAMI.2007.1067
