Next Article in Journal
Effective Heart Disease Prediction Using Machine Learning Techniques
Next Article in Special Issue
An Adaptive Deep Learning Neural Network Model to Enhance Machine-Learning-Based Classifiers for Intrusion Detection in Smart Grids
Previous Article in Journal
Efficient Online Engagement Analytics Algorithm Toolkit That Can Run on Edge
Previous Article in Special Issue
Detection of Cyberattacks and Anomalies in Cyber-Physical Systems: Approaches, Data Sources, Evaluation
 
 
Article
Peer-Review Record

From Iris Image to Embedded Code: System of Methods

Algorithms 2023, 16(2), 87; https://doi.org/10.3390/a16020087
by Ivan Matveev 1 and Ilia Safonov 2,*
Reviewer 1:
Reviewer 2:
Algorithms 2023, 16(2), 87; https://doi.org/10.3390/a16020087
Submission received: 15 December 2022 / Revised: 19 January 2023 / Accepted: 3 February 2023 / Published: 6 February 2023

Round 1

Reviewer 1 Report

In the considered manuscript, the authors propose algorithms ("the system/set of methods") in the field of cryptography: for generating a public key from both biometric data (iris image) and secret key (password). The authors provide a detailed description of their approach and seemingly well-grounded mathematical apparatus (even though I'm not an expert in Cybersecurity). However, the paper nearly completely lacks comparison with the state-of-the-art, which is in my opinion unacceptable for a paper whose claimed contribution is a new method.
So, I cannot recommend accepting the paper at the current stage and have to recommend major revisions. Some problems I saw with the paper are detailed below.

First, although the paper's topic is definitely relevant for the Algorithms journal, I have doubts whether it fits the Evolutionary Algorithms and Machine Learning section. I cannot see how the proposed approach is based on evolutionary/genetic algorithms, since it does not mention a fitness function, populations, etc. ML can be relevant only in a rather wider sense, and probably only for the segmentation method, even though I cannot quite find the learning even there. Neither the whole approach relies on methods that are commonly associated with Artificial Intelligence, so I would suggest that the editors consider the relevance of the manuscript for the "AI for Cybersecurity" Special Issue.
Also, I would recommend the authors to provide a background for their methods and better explain what family/field they relate to.

Second (and arguably the more important), the authors claim that their main contribution is as follows: "The set of methods allowing to build biometric cryptosystem based of iris images is presented. It contains three main parts: iris segmentation, biometric template generation, and the method of embedding/extracting the cryptographic key to/from biometric features."
However, neither of the three parts involves evaluation. The claim that "The system was tested on several databases of iris images. " seemingly relates to "6. Determining the threshold" and Table 1. However, this is rather far from an evaluation, as no alternative methods are considered.
Particularly, for iris segmentation (which is arguably the most relevant for ML) the novelty and the improved quality of the method, if any, needs to be demonstrated.
The overall framework of the methods proposed by the authors should be also compared to the alternatives using pre-formulated criteria.

Third (related to the previous point), the authors do not provide a comprehensive state-of-the-art review for any of the three parts of their methodology. For the segmentation this is limited to "Methods, algorithms and applications of iris biometrics are developing rapidly in recent years [4].", without any further background. I believe that the authors should more clearly demonstrate some limitations or disadvantages of alternative methods and justify the novelty of their approach.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The abstract section could be more impressive. In its current form, the Abstract is ambiguous. This section must provide a pertinent overview of the work. Also, quantitative values must be presented to highlight the work's novelty.

 

Introduction section: This section needs a hard improvement. What is the controversy of the study? The contribution and novelty need to be clarified. It is recommended that the authors deeply analyze the state of the art to find the novelty. Please discuss state of the art in this section. 

 

The objective of the work needs to be clarified.

 

Line 131: Please provide a flowchart of the method. 

 

The article does not present a Result section. 

 

Line 435: Please provide a comparison between the methods.

 

Bibliography: Please read the guide for authors before submitting a manuscript.

 

Line 86 can include the morphology tools as opening, gradient calculation, and other related issues. Please, consider these references to introduce these topics:

Hyperconnected openings codified in a max tree structure: an application for skull-stripping in brain mri t1; Differences in the visual performances of patients with strabismus, amblyopia, and healthy controls; Cortical activity at baseline and during light stimulation in patients with strabismus and amblyopia;

 

In general, the manuscript could be easier to follow. The novelty and contributions need to be clarified. If the article is a state-of-the-art review. The conclusion needs to show how new knowledge is generated. For this, is not possible to extend my recommendation for publication.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper presents a concept of iris image-based cryptographic key embedding. The paper is well-written and easy to read. It contains three major elements: iris processing and segmentation, binary template generation and embedding/extraction method. All of the elements are known and have been investigated previously. One can say there is nothing entirely new in the approach presented in the paper. So this is a compilation of well-known methods from a few scientific sub-areas. So the authors' motivation is unclear since many similar methods do exactly the same thing.

I would recommend rewriting the paper in order to state the most significant contribution since the iris detection, segmentation and post-processing (template construction) have already been done by many researchers, including Daugman, Wildes, Tisse, and Ferreira. What is more, the method presented in the paper is not the most successful one and has not even been compared with any others. Please refer to [1], [4], and [5]. Moreover, the error correction codes, pattern matching and information embedding (steganography) are also represented by many available works. The last thing, secret key embedding, is also nothing new. Please refer to [2] and [3]. In general, the paper lacks recent literature, e.g. related to deep learning.

 

[1] Radman, A., Jumari, K., Zainal, N. (2011). Iris Segmentation: A Review and Research Issues. In: Mohamad Zain, J., Wan Mohd, W.M.b., El-Qawasmeh, E. (eds) Software Engineering and Computer Systems. ICSECS 2011. Communications in Computer and Information Science, vol 179. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-22170-5_60

[2] H. M. Therar, E. A. Mohammed and A. J. Ali, "Biometric Signature based Public Key Security System," 2020 International Conference on Advanced Science and Engineering (ICOASE), 2020, pp. 1-6, doi: 10.1109/ICOASE51841.2020.9436615.

[3] Kumar, Mahesh & Prasad, Munaga & Raju, U.S.N.. (2020). BMIAE: Blockchain-based Multi-instance Iris Authentication using Additive ElGamal Homomorphic Encryption. IET Biometrics. 9. 10.1049/iet-bmt.2019.0169.

[4] Jasem Rahman Malgheet, Noridayu Bt Manshor, Lilly Suriani Affendey, "Iris Recognition Development Techniques: A Comprehensive Review", Complexity, vol. 2021, Article ID 6641247, 32 pages, 2021. https://doi.org/10.1155/2021/6641247

[5] Meng, Y.; Bao, T. Towards More Accurate and Complete Heterogeneous Iris Segmentation Using a Hybrid Deep Learning Approach. J. Imaging 2022, 8, 246. https://doi.org/10.3390/jimaging8090246

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I have read the authors' replies to my comments and to the comments made by the other reviewers, as well as the revised version of the manuscript. I'm glad to note that all the major issues were addressed:

- SOTA in Section 2,

- Intermediate evaluation in Section 4,

- Final evaluation and discussion in Section 10.

I also find the reply regarding the paper's topic relevance to the journal convincing enough.

So, I commend the authors for their work and recommend accepting the paper.

Reviewer 2 Report

The manuscript can be accepted

Reviewer 3 Report

I appreciate all the answers and corrections. Although I am not entirely convinced of the work's motivation and novelty, I am leaning towards publishing the paper in the current form.

Back to TopTop