All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
According to the comments of reviewers, after comprehensive consideration, your submission can be published.
[# PeerJ Staff Note - this decision was reviewed and approved by Vicente Alarcon-Aquino, a PeerJ Computer Science Section Editor covering this Section #]
I am satisfied with the modification of this paper.
I am satisfied with the modification of this paper.
I am satisfied with the modification of this paper.
The paper could be further enhanced by improving the descriptions of the formulas. For instance, in reference to Equation (7), the authors mention, "The results are too small". To provide a clearer mathematical context, it would be beneficial to define the value range of the variable. In this case, it could be stated as 'P(x|\theta) \in (0,1)'.
The proposed scheme is presented in great detail, and the language has been has been greatly improved.
The experimental design is logical and interesting.
The findings are neutral, and conclusions are also clearly stated.
All my comments have been well-addressed. This reviewer recommends acceptance of this manuscript.
Please make revisions to the paper based on the reviewer's comments.
The improvement of the paper and the authors' responses have addressede my concerns.
The improvement of the paper and the authors' responses have addressede my concerns.
The improvement of the paper and the authors' responses have addressede my concerns.
The proposed scheme is presented in great detail, but the language needs substantial improvement.
The experimental design is logical and interesting.
The findings are neutral, and if possible, more state-of- the-art models should be introduced for comparison.
1) Language needs great improvement. A large number of typos are observed in the current manuscript. For example, in paragraph two, Section “Related work”, “Sarier (2021) proposed a a new multimodal biometric authentication” should be “Sarier (2021) proposed a new multimodal biometric authentication”. In paragraph one, Section “Methodology”, “In the process of authenticating he user voice and face samples” should be “In the process of authenticating the user voice and face samples”. In paragraph one, Section “Face Recognition”, “making the face-recognition process more challenging. .” should be “making the face-recognition process more challenging.”. In addition, there are many sentences are confusing. Please double check the whole manuscript.
2) More state-of- the-art models should be introduced for comparison. The methods currently compared by the authors lack the achievements of the last two years, so the reviewer is concerned about the superiority of the proposed method.
3) With biometric authorization at the core of this research, some novel work needs to be mentioned. If possible, an overview of class-imbalance privacy-preserving federated learning for decentralized fault diagnosis with biometric authentication is needed.
4) Many abbreviations are defined repeatedly. For example, in Section "Gaussian Mixture Model", MFCC and GMM have been previously defined.
5) The quality of Figs. 3 and 6 could be improved.
Revise according to the comments of the reviewers.
[# PeerJ Staff Note: Please ensure that all review comments are addressed in a rebuttal letter and any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate. It is a common mistake to address reviewer questions in the rebuttal letter but not in the revised manuscript. If a reviewer raised a question then your readers will probably have the same question so you should ensure that the manuscript can stand alone without the rebuttal letter. Directions on how to prepare a rebuttal letter can be found at: https://peerj.com/benefits/academic-rebuttal-letters/ #]
I think the authors have summarized and motivated the problem really well. It is clear that they want to present a method for a multimodal biometric authentication system as using multiple modalities as input is better than a unimodal system.
They have referred to a lot of prior work in this domain too.
The method adopted by the authors has been described very clearly. However, I am not sure I understand how their method is different from the existing work?
Even in Table 2, they show that their method is performing better than 7 other related works. But, what exactly is causing this increase in performance.
So, it is important to analyze this performance with respect to their own method and other methods.
The dataset used is pretty big. Was the entire dataset used for producing these results? If not, then why?
Moreover, the dataset paper itself presented results on certain baselines using deep learning networks. I would believe using a deep learning network will perform much better on this task too. did the authors explore that?
Also, why have the results only been presented on 1 dataset? There are so many such audio-visual datasets that could also have been used for this task. This is important to report as this will speak to the generalizability of this proposed method.
- I think significant more experiments are required to show the validity of the work.
- Moreover an in-depth analysis on why this method is working better than the 7 other methods is required.
This paper is well-written in English and well-organized. The authors proposed an improved multimodal authentication system using FaceNet and GMM to deal with face and voice signals. The authors conducted experiments on a public dataset and their proposed method achieved good performance on the dataset. But there are still some text descriptions that can be improved:
1. In sentence 68, can the authors add a figure of the unimodal biometric system? It will be helpful for readers to capture the difference between the unimodal and multimodal systems.
2. In sentence 85-86, authors wrote 'The rest of paper.....', but only section 2 is involved in the following texts. Other sections should also be involved here.
3. In sentences 159 and 161, 'Abozaid et al.' is duplicated.
4. For Figures 2-4, some descriptions of the workflow of the frameworks may need to be added.
Some questions for the experiments:
1. In sentences 192 196, the authors wrote "alpha = 0.98...." and "alpha_0 = 0.53836". Are there any experimental results that can support these parameter selections?
2. In Table2, the authors show some existing methods' results. Do all of the methods conduct experiments in the same dataset? If not, I think only the experimental results within the same dataset can be shown here.
3. For Figure 7 roc curve, auc scores may need to be added in the figure legend for all curves.
No comment.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.