Data Matters: Rethinking the Data Distribution in Semi-Supervised Oriented SAR Ship Detection
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsDear Authors,
here, it follows my contributions to the improvement of this paper.
Minor review:
- TITLE: please, check the word "oriented". It is wrongly written.
- Figure 2: Where is the end of the process? Is it the Equation 10?
- Figure 7: Despite it being depicted in a line that saves space, one can see it and analyze it better if it could be depicted in two lines in a bigger size.
- Figure 8: Authors could enhance the visualization of some challenging cases depicted. Also, you could discuss some aspects of the distribution graph and the correspondent SAR image.
- Figure 11: Same comments as in Figure 8
Major review:
It is clear that the Authors made a big effort with a large data amount and the idea is interesting. However, the methodology is not clear. There are many secondary and tertiary procedures described. It makes the text understanding quite confusing. It is not clear where is the real scientific contribution. The authors made a technical work. However, the claimed scientific contribution seems to be weakly relevant.
Thus, the present work needs to be rewritten. The methodology must be clear and the Authors must emphasize the scientific contribution. Particularly, I consider that a bunch of procedures can solve a problem. However, it is not a guarantee that is a scientific study. The Authors must emphasize clearly the relation cause-effect for each step of the proposed methodology.
Moreover, the Authors must dedicate a Section describing in detail the dataset. In this current version, such a Section is mixed with experiments.
At last, I suggest that the Authors search for related works that focus on "annotation". It seems that it could have good simpler and effective solutions. Doing a comparison between them and yours will be very useful.
Anyway, I congratulate the Authors for the whole effort. Nonetheless, according to my point of view, the present work only could be accepted after major revision. I finalize my review encouraging the Authors to do it.
Comments on the Quality of English LanguageI would like to suggest to the Authors that each paragraph can be written with a clear purpose, which means, it would have a subject discussed with the reasoning well established. Many times, along current version of the manuscript, one can be confused with the lack of idea connections among them.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for Authors1.The author has chosen an innovative title, "DATA MATTERS," and has evaluated relevant data using various metrics throughout the paper. However, in lines 224-227 on page 7, the assessment of the importance of these metrics appears to be overly subjective. It would be beneficial to include a more thorough analysis.
2.There are relatively few comparative methods discussed. Additional methods should be considered. For instance, only FCOS is mentioned for single-stage methods, and only Faster R-CNN is mentioned for two-stage methods.
3.In the ablation study section, please include visualized results that address the issues of speckle noise, strong sea clutter, and high sidelobe levels as mentioned in the Introduction.
4.In the Related Works section, the author should provide their own understanding and summary of the related content rather than directly quoting from the original sources.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
