No-Reference Image Quality Assessment Based on a Multitask Image Restoration Network
Round 1
Reviewer 1 Report
to evaluate the quality based on the feature differences between reference and distorted image. the proposed method achieves excellent performance on both synthetically and 23 authentically distorted databases.
minor comments:
Table 5 can be elaborated in more details add some more figures for Comparison of pseudo-reference image qualit .
English needs to be improved.
Plag needs to be checked.
Comments for author File: 
Comments.pdf
English needs to be improved
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
They propose a no-reference image quality assessment method is proposed based on a multitask image restoration network. More detailed comments are given as follows:
1- Why they use U-Net architecture? What about other deep learning such as CNN and other models such as AlexNet, vgg, GoogLeNet, and etc.?
2- In related works section, the recent references such as 2022 and 2023 is very few. There are many works in 2022 therefore add some of them.
3- In Experimental assessment criteria. Add the citation. In addition what about RMSE Measure?
4- Add the future works in end of conclusion section.
5- The recent references ratio below 50% around 38%. Add more recent references.
6- Discuss the limitations of the proposed method.
7- In training phase, dataset partition is randomly or not? I suggest u to used K-Fold Cross Validation. And how many K-Fold used?
8- In experiments results, the evaluation (train-test round) must repeated for N round. I suggest to repeat for many rounds to ensure that the bias was minimized.
Minor editing of English language required
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Great paper!
I have just one comment: instead of using the word 'task', sometimes you can use 'aim', 'goal', 'issue', 'problem', etc.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
Dear authors,
I found your study to be highly interesting and comprehensive. I would like to provide a few minor suggestions for your consideration:
1. Regarding the paper's structure: If the editor approves of the current sectioning, then it can be considered acceptable. However, if necessary, you may need to revise the section headings and some of the content accordingly.
2. I strongly recommend utilising plots instead of tables for enhanced presentation, improved readability, and better understanding of the data.
3. I am curious about the rationale behind choosing ResNet50 specifically for this study. Have you explored whether employing ResNet101 or ResNeXt architectures could potentially improve the performance of your model?
4. It would be valuable to discuss the resource utilisation and processing time in comparison to other techniques. How does your approach fare in terms of these factors when compared to alternative methods?
These suggestions aim to further enhance the quality and impact of your research. Thank you for considering them.
Minor corrections have been suggested in the attached reviewed document to enhance readability. Please refer to the document for the specific suggestions and make the necessary adjustments accordingly.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
