Page 137 - 2024-Vol20-Issue2

133 | Fathi & Aziz

ACKNOWLEDGMENT

The authors thank the University of Mosul, especially the
Computer Engineering Department, for supporting this re-
search.

CONFLICT OF INTEREST

The authors have no conflict of relevant interest to this article.

Fig. 9. The structure of the proposed ANN REFERENCES

In future, we plan to apply deep transfer learning for feature [1] X. Wu, X. Feng, X. Cao, X. Xu, D. Hu, M. B. Lo´pez,
extraction as an alternative method to the handcrafted features and L. Liu, “Facial Kinship Verification: A Compre-
extraction method that we had applied. hensive Review and Outlook,” International Journal of
Computer Vision, vol. 130, no. 6, pp. 1494–1525, 2022.
Accuracy = Number o f Correct Predictions *100% (3)
Total Number o Predictions [2] S. Albalawi, L. Alshahrani, N. Albalawi, R. Kilabi, and
A. Alhakamy, “A Comprehensive Overview on Biomet-
TABLE III. ric Authentication Systems using Artificial Intelligence
SUMMARY OF PROPOSED MODEL RESULTS Techniques,” Int J Adv Comput Sci Appl, vol. 13, no. 4,
pp. 782–791, 2022.
Number of training images 63 images
Number of testing images 16 images [3] K. P?rihodova´ and M. Hub, “Biometric Privacy Through
Hand Geometry-a Survey,” in 2019 International Con-
Train Accuracy 0.9677 ference on Information and Digital Technologies (IDT),
Loss 0.1428 pp. 395–401, IEEE, 2019.
0.9375
Prediction accuracy [4] M. Bac?a, P. Grd, and T. Fotak, “Basic Principles and
Training time 82s Trends in Hand Geometry and Hand Shape Biometrics,”
Prediction time 1s New Trends and Developments in Biometrics, pp. 77–99,
2012.
V. CONCLUSION
[5] T. Sabhanayagam, V. P. Venkatesan, and K. Senthama-
In this work, we created a database for studying the poten- raikannan, “A Comprehensive Survey on Various Bio-
tial of employing hand biometric for kinship verification and metric Systems,” International Journal of Applied Engi-
have offered a set of helpful metadata of the proposed dataset, neering Research, vol. 13, no. 5, pp. 2276–2297, 2018.
which is still being built and is expected to grow. Additionally,
we used the ANN model for multiclassification problem to [6] W. Wang, S. You, S. Karaoglu, and T. Gevers, “A Sur-
test the suggested dataset. The results of this novel approach vey on Kinship Verification,” Neurocomputing, vol. 525,
demonstrated that the hand possesses biometric characteristics pp. 1–28, 2023.
that may be used to establish kinship, and that the suggested
method is a promising way as a kinship indicator. As a re- [7] M. M. T. Al Mossawy and L. E. George, “A Digital Sig-
placement for the manual feature extraction method, we used nature System Based on Hand Geometry-Survey: Basic
in the past, deep transfer learning is a method we intend to Components of Hand-based Biometric System,” Wasit
use in the future. Journal of Computer and Mathematics Science, vol. 1,
no. 1, pp. 1–14, 2022.

[8] P. Varchol and D. Levicky, “Using of Hand Geometry in
Biometric Security Systems,” Radio engineering Prague,
vol. 16, no. 4, p. 82, 2007.

[9] S. Angadi and S. Hatture, “ Hand Geometry Based User
Identification Using Minimal Edge Connected Hand Im-
age Graph,” IET computer vision, vol. 12, no. 5, pp. 744–
752, 2018.