Robust Authentication System with Privacy Preservation of Biometrics

Table 1

Comparison of state-of-the-art multi-biometric authentication systems with different biometrics.

Reference	Methodology	Performance/accuracy	Irreversibility	Unlinkability	Renewability	Limitations
Gomez-Barrero 2017 [2]	Protection of multi-biometric templates via homomorphic encryption	High accuracy rates, reaching EERs as low as 0.12%, and requiring protected templates comprising 200 KB	√	√	---	Renewability needs to attain
Dong, J., Meng, X., Chen, M. and Wang, Z., 2017 [6]	For multimodal biometric identification, pattern defense based on DNA coding is required	ERR of the presented algorithm is 3.6%	---	---	√	Unlinkability and irreversibility standards are not taken into consideration
Ahmad, M., Mohamad, N., Md Isa, M., Ngadiran, R. and Darsono, A., 2017 [7]	Fusion of DCT image low-frequency coefficients for multimodal biometrics using face and palm print	The recognition rate is 95%	---	---	---	Not focused on ISO standards
Jagadiswary, D. and Saraswady, D., 2016 [8]	Using fused multimodal biometrics for biometric authentication	Performance of multimodal biometrics with RSA has GAR of 95.3% and FAR of 0.01%	---	---	---	-
Nandhinipreetha, A. and Radha, N., 2016 [9]	Visual cryptography validates multimodal biometric templates such as the finger vein and signature	The system is increased with a low FAR of 2%, FRR of 1.3%, and accuracy of 98.2%	---	---	---	-
M. K., Prasad, M. V. N. K. and Raju, U. S. N. (2020) [21],	Using completely homomorphic encryption, privacy-preserving iris authentication	(EER = 0.19%, 0.39%, 0.99%, and 0.28% for CASIA-V 1.0, CASIA-V3-Interval, IITD, and SDUMLA-HMT iris databases)	√	√	√	All ISO standards are satisfied
Gupta, A. et al. (2014) [37]	We can combine face, palm vein, and palm print modalities at the feature level using the discrete cosine transform	This methodology has 0% FAR with 100% GAR with Canberra distance	---	---	---	---
Peng, J. et al. (2014) [17]	Multi-biometric fingerprint cryptosystems: a fusion approach and template security	Three fusion strategies, feature-level, score-level, and decision-level mergers with the corresponding template protection technique, are proposed as multi-biometric finger cryptosystems	√	---	---	Huge memory required
Vishi, K. and Mavroeidis, V. (2018) [18]	A comparison of methods for fusing fingerprint and finger vein biometric data at the score level	Evaluated permutations of score convergence and ways of engaging using two modalities	---	---	---	Privacy and security concerns are not taken into consideration
Rathgeb, C. et al. (2016) [19]	Improved multi-biometric fuzzy iris vault that is unlinkable	An unlinkable multi-instance iris biometric cryptographic algorithms were obtained with the presented fuzzy vault system	---	---	---	Privacy and security concerns are not taken into consideration
Stokkenes, M. et al. (2016) [20]	Protection against multiple biometric templates—an study of the security implications of binarized statistical characteristics used in bloom filters for cellphones	Presented system based on bloom filters and binary statistical image features (BSIFs)	√	√	---	Renewability is not being attained
Morampudi, M. K., Prasad, M. V. N. K., &Raju, U. S. N. (2020) [21]	Using completely homomorphic encryption to protect the privacy of iris authentication	Introduced a method of privacy policy-preserving iris verification employing completely homomorphic encryption	√	√	√	All ISO standards are taken into consideration
Kirchgasser, S. et al. (2020) [22]	Multiple face biometrics in the visible and near-infrared light domain are protected using templates	Introduced template security mechanisms, including cancellable biometrics	---	√	---	One ISO standard is attained
Nafea, O. et al. (2016) [38]	Watermarking is used to secure hybrid multi-biometric templates	The scheme is based on a watermarking process, a shuffling process, and a Hadamard matrix	---	---	---	---
Gomez-Barrero, M. et al. (2018) [23]	Protection of several biometric templates using bloom filters	Presented an approach for predicting the principal components of the multi-biometric protection template focusing on bloom filters	√	√	---	Renewability is not being attained
Chang, D. et al. (2020) [24]	A cancelable multi-biometric method is possible using a fuzzy separator and innovative bitwise encoding	Presented a cancellable multi-biometric authentication of users in which a unique bitwise encryption method was used	√	---	---	All ISO standards not attained
Drozdowski, P. et al. (2018) [24]	Indexing iris codes using cancelable bloom filter-based exploration assemblies while maintaining privacy	Presented the first-ever method for archiving protected iris templates, which essentially seeks to minimize tasks in computationally intensive verification mode	√	---	---	All ISO standards not attained
S. S. et al. (2020) [29]	Biometric identification solution that is both robust and safe	Presented a technique in which a strong 3-dimensional template was developed using the fingerprint of the individual	---	---	---	Revocability, diversity is attained
Jutta H¨ammerle-Uhl.et.al (2009) [34]	Cancelable iris biometrics using block remapping and image warping	Presented cancelable biometric protection technique where two classical transformations, block remapping and texture warping, were applied to iris textures	√	---	---	One ISO standard is attained
LUIGI CATUOGNO et. al. (2016) [35]	An enterprise rights management system for on-the-field maintenance facilities	A secure multi-biometric authentication scheme is used to identify operators while guaranteeing system usability and user privacy	---	---	---	Security is attained but not all ISO standards
Sharat Chikkerur et.al. (2008) [36]	Generating registration-free cancelable fingerprint templates	A secure, registration-free construction of cancelable fingerprint templates is proposed. New fingerprint representation and from that cancelable biometrics constructed	√	---	---	Non-invertible variations achieved