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.
Dear authors, we are pleased to verify that you meet the reviewer's valuable feedback to improve your research.
Thank you for considering PeerJ Computer Science and submitting your work.
[# PeerJ Staff Note - this decision was reviewed and approved by Xiangjie Kong, a PeerJ Section Editor covering this Section #]
I have checked the revised manuscript and confirm that the authors have implemented all requested changes. The manuscript has been significantly improved.
The authors could improve the visibility of Figures 3, 4 and 5. It is very difficult to distinguish different datasets (CASIA and CelebA) based only on the slightly changed colour of the borders. This problem is particularly pronounced in monochrome printing, which most of us practice. They should change the line type for one data set. Also, Figure 3 could have a more detailed caption of experimental results.
no comment
no comment
The authors have revised the paper to address the comments of reviewers. It appears good now.
Authors have reported and explain the results sufficiently.
Authors have reported and explain the results sufficiently.
N/A
1. Background should be after Introduction.
My previous feedback has been addressed, and everything looks good now.
My previous feedback has been addressed, and everything looks good now.
Dear authors,
You are advised to critically respond to all comments point by point when preparing a new version of the manuscript and while preparing for the rebuttal letter. Please address all the comments/suggestions provided by the reviewers.
Kind regards,
PCoelho
no comment
To show the robustness of your approach, the work can be tested on other popular datasets.
no comment
Future prospects of the proposed approach should be discussed.
While the paper contains valuable research, the overall quality of the manuscript is significantly diminished due to numerous issues related to writing style and grammatical accuracy. Firstly, the writing style is often convoluted, making it challenging for the reader to follow the arguments and understand the key points. Repetitive wording in a sentence (e.g. adversarial) should be revised for clarity and conciseness. Sentences tend to be overly complex, which not only obscures the main ideas but also leads to confusion. A more concise and straightforward style would greatly enhance the clarity and readability of the paper. Moreover, there are frequent grammatical errors throughout the manuscript. These include incorrect verb tenses, subject-verb agreement issues, and improper use of punctuation.
Here are a few examples:
• There are formatting issues, right in the first lines of the paper (abstract), where there are missing spaces after periods and before parentheses.
• For instance, in line 40, incorrect subject-pronoun agreement: "The digital content themselves obtain characteristic information …" should be "The digital content itself obtains characteristic information …"
• The same title is found in lines 401 and 428?
It would be beneficial for the authors to add a paragraph at the end of the introduction section to describe the paper's structure and sections.
Authors need to confirm that all acronyms are defined before being used for the first time.
The paper would benefit from a thorough proofreading and editing process to address these issues. Seeking assistance from a professional editor or utilizing advanced grammar-checking tools could be beneficial. Improving these aspects will help in better conveying the study’s findings and enhancing its impact on the scientific community.
No comment at this point
No comment at this point
• I would suggest authors to include a comparative table of prior works considering different parameters.
• At the end of the introduction, it would be nice to include the structure of the paper.
• More scientific reasoning should be added in the experimental results' explanations.
• Improve your discussion of results through the cause and effect of the modeling parameters, setting values, and model hyper-parameters.
• Models’ validation section needs to be enriched with more details and dataset.
• Authors missed many recent references such as in the reference section, e.g.,
Akhtar Z. Deepfakes generation and detection: a short survey. Journal of Imaging. 2023 Jan 13;9(1):18.
Abbas F, Taeihagh A. Unmasking deepfakes: A systematic review of deepfake detection and generation techniques using artificial intelligence. Expert Systems With Applications. 2024 May 18:124260.
**PeerJ Staff Note:** It is PeerJ policy that additional references suggested during the peer-review process should only be included if the authors are in agreement that they are relevant and useful.
1. The background section could be separated from the current introduction section to improve clarity.
2. The plots for experimental results are difficult to read. The histograms and legends are not presented in a straightforward way.
1. The algorithm would benefit from further discussion on reducing the generation time of adversarial examples.
2. While the DADFI algorithm demonstrates good transferability across different deepfake models, its performance could potentially be enhanced by exploring additional techniques or modifications to the training process.
3. The aggressiveness of the adversarial samples warrants further evaluation. Investigating methods to maintain or improve aggressiveness without requiring the fine-tuning step could be a valuable area for enhancement.
1. The limitations and discussion of future work appear to be missing.
2. Using a wider range of deepfake models and datasets would be better to ensure the algorithm's robustness and generalizability.
N/A
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.