TY  - JOUR
AU  - Gnjatović, Milan
AU  - Maček, Nemanja
AU  - Saračević, Muzafer
AU  - Adamović, Saša
AU  - Joksimović, Dušan
AU  - Karabašević, Darjan
PY  - 2023
UR  - http://jakov.kpu.edu.rs/handle/123456789/1514
AB  - This paper introduces a heuristic for multiple sequence alignment aimed at improving
real-time object recognition in short video streams with uncertainties. It builds upon the idea of
the progressive alignment but is cognitively economical to the extent that the underlying edit dis-
tance approach is adapted to account for human working memory limitations. Thus, the proposed
heuristic procedure has a reduced computational complexity compared to optimal multiple sequence
alignment. On the other hand, its relevance was experimentally confirmed. An extrinsic evaluation
conducted in real-life settings demonstrated a significant improvement in number recognition accu-
racy in short video streams under uncertainties caused by noise and incompleteness. The second line
of evaluation demonstrated that the proposed heuristic outperforms humans in the post-processing
of recognition hypotheses. This indicates that it may be combined with state-of-the-art machine
learning approaches, which are typically not tailored to the task of object sequence recognition from a
limited number of frames of incomplete data recorded in a dynamic scene situation
PB  - Basel, Switzerland : MDPI
T2  - Axioms
T1  - Cognitively Economical Heuristic for Multiple Sequence Alignment under Uncertainties
VL  - 12
IS  - 1
SP  - 3
DO  - 10.3390/axioms12010003
ER  - 

@article{
author = "Gnjatović, Milan and Maček, Nemanja and Saračević, Muzafer and Adamović, Saša and Joksimović, Dušan and Karabašević, Darjan",
year = "2023",
abstract = "This paper introduces a heuristic for multiple sequence alignment aimed at improving
real-time object recognition in short video streams with uncertainties. It builds upon the idea of
the progressive alignment but is cognitively economical to the extent that the underlying edit dis-
tance approach is adapted to account for human working memory limitations. Thus, the proposed
heuristic procedure has a reduced computational complexity compared to optimal multiple sequence
alignment. On the other hand, its relevance was experimentally confirmed. An extrinsic evaluation
conducted in real-life settings demonstrated a significant improvement in number recognition accu-
racy in short video streams under uncertainties caused by noise and incompleteness. The second line
of evaluation demonstrated that the proposed heuristic outperforms humans in the post-processing
of recognition hypotheses. This indicates that it may be combined with state-of-the-art machine
learning approaches, which are typically not tailored to the task of object sequence recognition from a
limited number of frames of incomplete data recorded in a dynamic scene situation",
publisher = "Basel, Switzerland : MDPI",
journal = "Axioms",
title = "Cognitively Economical Heuristic for Multiple Sequence Alignment under Uncertainties",
volume = "12",
number = "1",
pages = "3",
doi = "10.3390/axioms12010003"
}

Gnjatović, M., Maček, N., Saračević, M., Adamović, S., Joksimović, D.,& Karabašević, D.. (2023). Cognitively Economical Heuristic for Multiple Sequence Alignment under Uncertainties. in Axioms
Basel, Switzerland : MDPI., 12(1), 3.
https://doi.org/10.3390/axioms12010003

Gnjatović M, Maček N, Saračević M, Adamović S, Joksimović D, Karabašević D. Cognitively Economical Heuristic for Multiple Sequence Alignment under Uncertainties. in Axioms. 2023;12(1):3.
doi:10.3390/axioms12010003 .

Gnjatović, Milan, Maček, Nemanja, Saračević, Muzafer, Adamović, Saša, Joksimović, Dušan, Karabašević, Darjan, "Cognitively Economical Heuristic for Multiple Sequence Alignment under Uncertainties" in Axioms, 12, no. 1 (2023):3,
https://doi.org/10.3390/axioms12010003 . .

TY  - JOUR
AU  - Barzut, Srđan
AU  - Milosavljević, Milan
AU  - Adamović, Saša
AU  - Saračević, Muzafer
AU  - Maček, Nemanja
AU  - Gnjatović, Milan
PY  - 2021
UR  - http://jakov.kpu.edu.rs/handle/123456789/1511
AB  - Modern access controls employ biometrics as a means of authentication to a great extent.
For example, biometrics is used as an authentication mechanism implemented on commercial devices
such as smartphones and laptops. This paper presents a fingerprint biometric cryptosystem based on
the fuzzy commitment scheme and convolutional neural networks. One of its main contributions
is a novel approach to automatic discretization of fingerprint texture descriptors, entirely based on
a convolutional neural network, and designed to generate fixed-length templates. By converting
templates into the binary domain, we developed the biometric cryptosystem that can be used in
key-release systems or as a template protection mechanism in fingerprint matching biometric systems.
The problem of biometric data variability is marginalized by applying the secure block-level Bose–
Chaudhuri–Hocquenghem error correction codes, resistant to statistical-based attacks. The evaluation
shows significant performance gains when compared to other texture-based fingerprint matching
and biometric cryptosystems.
PB  - Basel, Switzerland : MDPI
T2  - Mathematics
T1  - A Novel Fingerprint Biometric Cryptosystem Based on Convolutional Neural Networks
VL  - 9
IS  - 7
SP  - 730
DO  - 10.3390/math9070730
ER  - 

@article{
author = "Barzut, Srđan and Milosavljević, Milan and Adamović, Saša and Saračević, Muzafer and Maček, Nemanja and Gnjatović, Milan",
year = "2021",
abstract = "Modern access controls employ biometrics as a means of authentication to a great extent.
For example, biometrics is used as an authentication mechanism implemented on commercial devices
such as smartphones and laptops. This paper presents a fingerprint biometric cryptosystem based on
the fuzzy commitment scheme and convolutional neural networks. One of its main contributions
is a novel approach to automatic discretization of fingerprint texture descriptors, entirely based on
a convolutional neural network, and designed to generate fixed-length templates. By converting
templates into the binary domain, we developed the biometric cryptosystem that can be used in
key-release systems or as a template protection mechanism in fingerprint matching biometric systems.
The problem of biometric data variability is marginalized by applying the secure block-level Bose–
Chaudhuri–Hocquenghem error correction codes, resistant to statistical-based attacks. The evaluation
shows significant performance gains when compared to other texture-based fingerprint matching
and biometric cryptosystems.",
publisher = "Basel, Switzerland : MDPI",
journal = "Mathematics",
title = "A Novel Fingerprint Biometric Cryptosystem Based on Convolutional Neural Networks",
volume = "9",
number = "7",
pages = "730",
doi = "10.3390/math9070730"
}

Barzut, S., Milosavljević, M., Adamović, S., Saračević, M., Maček, N.,& Gnjatović, M.. (2021). A Novel Fingerprint Biometric Cryptosystem Based on Convolutional Neural Networks. in Mathematics
Basel, Switzerland : MDPI., 9(7), 730.
https://doi.org/10.3390/math9070730

Barzut S, Milosavljević M, Adamović S, Saračević M, Maček N, Gnjatović M. A Novel Fingerprint Biometric Cryptosystem Based on Convolutional Neural Networks. in Mathematics. 2021;9(7):730.
doi:10.3390/math9070730 .

Barzut, Srđan, Milosavljević, Milan, Adamović, Saša, Saračević, Muzafer, Maček, Nemanja, Gnjatović, Milan, "A Novel Fingerprint Biometric Cryptosystem Based on Convolutional Neural Networks" in Mathematics, 9, no. 7 (2021):730,
https://doi.org/10.3390/math9070730 . .

TY  - JOUR
AU  - Gnjatović, Milan
AU  - Maček, Nemanja
AU  - Adamović, Saša
PY  - 2020
UR  - http://jakov.kpu.edu.rs/handle/123456789/1516
AB  - This paper introduces a novel approach to human-machine collaborative
learning that allows for the chronically missing human learnability in the context of
supervised machine learning. The basic tenet of this approach is the refinement of a human
designed software model through the iterative learning loop. Each iteration of the loop
consists of two phases: (i) automatic data-driven parameter adjustment, performed by
means of stochastic greedy local search, and (ii) human-driven model adjustment based on
insights gained in the previous phase. The proposed approach is demonstrated through a
real-life study of automatic electricity meter reading in the presence of noise. Thus, a
cognitively-inspired non-connectionist approach to digit detection and recognition is
introduced, which is subject to refinement through the iterative process of human-machine
cooperation. The prototype system is evaluated with respect to the recognition accuracy
(with the highest digit recognition accuracy of 94%), and also discussed with respect to the
storage requirements, generalizability, utilized contextual information, and efficiency.
PB  - Budapest : Óbuda University
T2  - Acta Polytechnica Hungarica
T1  - Putting Humans Back in the Loop: A Study in Human-Machine Cooperative Learning
VL  - 17
IS  - 2
SP  - 191
EP  - 210
DO  - 10.12700/APH.17.2.2020.2.11
ER  - 

@article{
author = "Gnjatović, Milan and Maček, Nemanja and Adamović, Saša",
year = "2020",
abstract = "This paper introduces a novel approach to human-machine collaborative
learning that allows for the chronically missing human learnability in the context of
supervised machine learning. The basic tenet of this approach is the refinement of a human
designed software model through the iterative learning loop. Each iteration of the loop
consists of two phases: (i) automatic data-driven parameter adjustment, performed by
means of stochastic greedy local search, and (ii) human-driven model adjustment based on
insights gained in the previous phase. The proposed approach is demonstrated through a
real-life study of automatic electricity meter reading in the presence of noise. Thus, a
cognitively-inspired non-connectionist approach to digit detection and recognition is
introduced, which is subject to refinement through the iterative process of human-machine
cooperation. The prototype system is evaluated with respect to the recognition accuracy
(with the highest digit recognition accuracy of 94%), and also discussed with respect to the
storage requirements, generalizability, utilized contextual information, and efficiency.",
publisher = "Budapest : Óbuda University",
journal = "Acta Polytechnica Hungarica",
title = "Putting Humans Back in the Loop: A Study in Human-Machine Cooperative Learning",
volume = "17",
number = "2",
pages = "191-210",
doi = "10.12700/APH.17.2.2020.2.11"
}

Gnjatović, M., Maček, N.,& Adamović, S.. (2020). Putting Humans Back in the Loop: A Study in Human-Machine Cooperative Learning. in Acta Polytechnica Hungarica
Budapest : Óbuda University., 17(2), 191-210.
https://doi.org/10.12700/APH.17.2.2020.2.11

Gnjatović M, Maček N, Adamović S. Putting Humans Back in the Loop: A Study in Human-Machine Cooperative Learning. in Acta Polytechnica Hungarica. 2020;17(2):191-210.
doi:10.12700/APH.17.2.2020.2.11 .

Gnjatović, Milan, Maček, Nemanja, Adamović, Saša, "Putting Humans Back in the Loop: A Study in Human-Machine Cooperative Learning" in Acta Polytechnica Hungarica, 17, no. 2 (2020):191-210,
https://doi.org/10.12700/APH.17.2.2020.2.11 . .

TY  - JOUR
AU  - Maček, Nemanja
AU  - Adamović, Saša
AU  - Milosavljević, Milan
AU  - Jovanović, Miloš
AU  - Gnjatović, Milan
AU  - Trenkić, Branimir
PY  - 2019
UR  - http://jakov.kpu.edu.rs/handle/123456789/1517
AB  - This paper presents an approach to designing secure modular authentication
framework based on iris biometrics and its’ implementation into mobile banking scenario.
The system consists of multiple clients and an authentication server. Client, a smartphone
with accompanying application, is used to capture biometrics, manage auxiliary data and
create and store encrypted cancelable templates. Bank’s authentication server manages
encryption keys and provides the template verification service. Proposed system keeps
biometric templates encrypted or at least cancelable during all stages of storage,
transmission and verification. As templates are stored on clients in encrypted form and
decryption keys reside on bank's authentication server, original plaintext templates are
unavailable to an adversary if the phone gets lost or stolen. The system employs public key
cryptography and pseudorandom number generator on small-sized templates, thus not
suffering from severe computational costs like systems that employ homomorphic
encryption. System is also general, as it does do not depend on specific cryptographic
algorithms. Having in mind that modern smartphones have iris scanners or at least high-
quality front cameras, and that no severe computational drawbacks exist, one may
conclude that the proposed authentication framework is highly applicable in mobile
banking authentication.
PB  - Budapest : Óbuda University
T2  - Acta Polytechnica Hungarica
T1  - Mobile Banking Authentication Based on Cryptographically Secured Iris Biometrics
VL  - 16
IS  - 1
SP  - 45
EP  - 62
DO  - 10.12700/APH.16.1.2019.1.3
ER  - 

@article{
author = "Maček, Nemanja and Adamović, Saša and Milosavljević, Milan and Jovanović, Miloš and Gnjatović, Milan and Trenkić, Branimir",
year = "2019",
abstract = "This paper presents an approach to designing secure modular authentication
framework based on iris biometrics and its’ implementation into mobile banking scenario.
The system consists of multiple clients and an authentication server. Client, a smartphone
with accompanying application, is used to capture biometrics, manage auxiliary data and
create and store encrypted cancelable templates. Bank’s authentication server manages
encryption keys and provides the template verification service. Proposed system keeps
biometric templates encrypted or at least cancelable during all stages of storage,
transmission and verification. As templates are stored on clients in encrypted form and
decryption keys reside on bank's authentication server, original plaintext templates are
unavailable to an adversary if the phone gets lost or stolen. The system employs public key
cryptography and pseudorandom number generator on small-sized templates, thus not
suffering from severe computational costs like systems that employ homomorphic
encryption. System is also general, as it does do not depend on specific cryptographic
algorithms. Having in mind that modern smartphones have iris scanners or at least high-
quality front cameras, and that no severe computational drawbacks exist, one may
conclude that the proposed authentication framework is highly applicable in mobile
banking authentication.",
publisher = "Budapest : Óbuda University",
journal = "Acta Polytechnica Hungarica",
title = "Mobile Banking Authentication Based on Cryptographically Secured Iris Biometrics",
volume = "16",
number = "1",
pages = "45-62",
doi = "10.12700/APH.16.1.2019.1.3"
}

Maček, N., Adamović, S., Milosavljević, M., Jovanović, M., Gnjatović, M.,& Trenkić, B.. (2019). Mobile Banking Authentication Based on Cryptographically Secured Iris Biometrics. in Acta Polytechnica Hungarica
Budapest : Óbuda University., 16(1), 45-62.
https://doi.org/10.12700/APH.16.1.2019.1.3

Maček N, Adamović S, Milosavljević M, Jovanović M, Gnjatović M, Trenkić B. Mobile Banking Authentication Based on Cryptographically Secured Iris Biometrics. in Acta Polytechnica Hungarica. 2019;16(1):45-62.
doi:10.12700/APH.16.1.2019.1.3 .

Maček, Nemanja, Adamović, Saša, Milosavljević, Milan, Jovanović, Miloš, Gnjatović, Milan, Trenkić, Branimir, "Mobile Banking Authentication Based on Cryptographically Secured Iris Biometrics" in Acta Polytechnica Hungarica, 16, no. 1 (2019):45-62,
https://doi.org/10.12700/APH.16.1.2019.1.3 . .