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,
One of the preceding reviewers declined to undertake a review of the revised paper. Conversely, another reviewer accepted the paper in its current form. It is evident that the paper has undergone substantial improvement and is now deemed suitable for publication.
Best wishes,
[# PeerJ Staff Note - this decision was reviewed and approved by Xiangjie Kong, a PeerJ Section Editor covering this Section #]
All comments properly entertained just
All point properly entertained
In this research paper all comments properly entertained and oky for publication from my side
All Comments properly entertained
Dear Authors,
Reviewers have now commented on your article. We do encourage you to address the concerns and criticisms of the reviewers with respect to reporting, experimental design, and validity of the findings and resubmit your article once you have updated it accordingly.
Best wishes,
This paper presents a solid foundation; however, there are some critical issues that need to be addressed. While these issues are not highly significant, resolving them is essential to strengthen the overall quality of the paper. Therefore, I have some queries related to the BiLSTM model architecture and training strategy that need clarification to overcome these weaknesses.
1. What is the architecture design and training strategy used in the BiLSTM model for [specific task or dataset].
2. Why was the BiLSTM model chosen over other models for this task.
3. Explain the loss function used in the model?
4.What is the role of the optimization mechanism in the training and performance of the BiLSTM model?
5. Explain detection optimizing strategies.
6.Why more fake video images used than real images in the study.
7. Describe the working LSTM and CRF.
8. Explain Fig 4.
9. Explain General framework of the model
10.Draw a Comparison graph.
How does the use of BiLSTM contribute to the validity and reliability of the findings in this research on fast detection of deep fake face videos for real-time applications.
Your research on the effectiveness and optimization of BiLSTM-based fast detection of deep fake face videos for real-time applications is promising. However, it is essential to clarify how the use of the BiLSTM model contributes to the validity and reliability of the findings. Providing detailed insights into why BiLSTM was chosen and how it enhances the detection process will strengthen the overall credibility of your work.
The study is well structured theoretically. The results are presented in detail with figures and tables.
The study is written in clear and understandable English.
The “Related Work” section provides an overview of the current state of the literature, but could be enriched with a more in-depth analysis and methodological comparisons. The methodologies used in these studies and the results obtained could be discussed in more detail. For example, more information could be given about which datasets were used for each method, the success rates achieved, the challenges and advantages. This would provide readers with a clearer understanding of the context and effectiveness of the methods. The shortcomings and gaps in existing studies could be analyzed in more detail. More comparisons could be made regarding the effectiveness of other deep learning methods other than BiLSTM in deepfake face video detection. For example, the success of alternatives such as CNN, GAN-based approaches, Transformer-based models and what advantages they offer compared to BiLSTM could be discussed.
Abbreviations should be explained only once in the paper. (For example: "bidirectional long short-term memory network (BiLSTM)". In subsequent sections, only the abbreviation should be used and the expansion should not be included each time.
It is insufficient to compare with Xception and a BiLSTM model without CRF. To show how effective their approach is, the authors should compare it to other state-of-the-art deepfake detection methods.
Precision and recall metrics should also be given for experimental results. In this way, the strengths and weaknesses of the model will be revealed more clearly.
Although the paper emphasizes real-time applications, it doesn't provide concrete metrics like frames per second (FPS) or latency to demonstrate the model's real-time capabilities.
The authors claim high efficiency and accuracy across multiple datasets, but only FaceForensics++ is mentioned. Experiments on other datasets are needed to support this claim.
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.