Blind Edge-Retention Indicator for Assessing the Quality of Filtered (Pol)SAR Images Based on a Ratio Gradient Operator and Confidence Interval Estimation
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis paper addresses the issue that the edge-preservation capability of PolSAR filters cannot be accurately assessed in the absence of reference images. A blind assessment method based on the SAR Ratio Gradient Operator and Confidence Interval Estimation is proposed. The authors evaluated the edge-preservation capability of different filtering algorithms on four SAR datasets.
In my opinion, the paper presents a novel approach for assessing edge-preservation capability, but the experimental validation and analysis of the proposed method incomplete. There are several unclear points in the paper, and some comments and suggestions are as follows:
(1) Page 5, line 36. Equation (11) is the most important equation in the paper. For the readers' better understanding, the authors should provide a more detailed explanation.
(2) Page 5, line 231. Equations (16) and (17) specify the confidence interval as [0.05, 0.9]. Whether this confidence interval is reasonable, it is recommended that the authors provide a reasonable criterion for determining the confidence interval based on a large number of experiments.
(3) Page 5, line 266. Regarding the RGO-BAI model proposed by the authors, it is suggested that the authors simulate a simple noise-free image, introduce multiplicative noise, and then verify whether the proposed method exhibits consistency across different filtering results.
(4) Page 8, line 297. In the experimental and results analysis section, it is hoped that the authors can analyze how the ratio images represent the edge preservation capability of SAR images in real data, and whether the presentation of the results in ratio images can be improved.
(5) With the improvement in resolution, the size of SAR images typically reaches 20,000 × 20,000 pixels or more. It is recommended that the authors expand the experimental content and analyze the computational efficiency of the proposed method.
Author Response
Dear reviewer, please refer to the attached file.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Comments and Suggestions for AuthorsGood work in SAR Speckle Reduction Assessment. Please solve the following comments with a minor revision and to be reviewed again.
1. Give more information of the three real-word SAR images in Abstract.
2. Give the link of source code and sample datasets in Abstract.
3. The statement of contributions at the end of Sec. 1 should be enriched more, also aiming at better highlighting the characteristics of this paper.
4. In Sec. 3.2, the authors are suggested to provide the framework figure of the proposed indicator.
5. In addition to the comparison in accuracy, the comparison with other models can be expanded to inference speed.
6. In Table 1, which assessment indictors are designed for SAR despeckling and which comparative methods are specifically designed for PolSAR despeckling?
7. In the further, the authors should consider adopt speckle reduction methods for SAR image interpretation works, i.e., SAR object classification, and SAR object detection, such as: HOG-ShipCLSNet: A novel deep learning network with hog feature fusion for sar ship classification, Lite-yolov5: a lightweight deep learning detector for on-board ship detection in large-scene sentinel-1 sar images, Scale in scale for sar ship instance segmentation, A group-wise feature enhancement-and-fusion network with dual-polarization feature enrichment for sar ship detection.
8. IMHO, the Conclusion should be re-written to 1) explicitly describe the essential features/advantages of the review that other reviews do not have, and 2) describe the limitation(s) of the review.
9. The English should be improved greatly.
Comments on the Quality of English LanguageModerate editing of English language required.
Author Response
Dear reviewer, please refer to the attached file.
Author Response File: Author Response.docx
Reviewer 3 Report
Comments and Suggestions for AuthorsDear authors,
I have carefully reviewed your work and find it to be a valuable contribution to the field of assessing the quality of despeckling SAR images. Your proposed evaluation index, which considers the gradient operator ratio and confidence interval estimation, shows promise for practical applications, particularly in accurately assessing the filtering results of SAR images. Overall, the structure of the manuscript is well-organized, and the content is well-written. It is evident that you have conducted thorough research, as demonstrated by the final results. The performance and effectiveness of your proposed framework are commendable. I only have some minor comments on this work, which may improve the quality of the manuscript:
1. Title: The current title is fine, but it seems more like an indicator to evaluate the filtering methods. From my perspective, the proposed indicator can be more useful for assessing the quality of filtered SAR images than as an indicator to evaluate the method itself.
2. Conclusion: The conclusion part should be revised carefully to provide a more comprehensive summary of the paper and its findings.
3. Additional Experiments: It would be beneficial to include a comparison of computation time with traditional indices. This comparison can help illustrate the advantages and disadvantages of your proposed indicator to the readers.
4. Typos: such as a typo on line 460. Please review and correct the typos thorough out the whole paper.
Comments on the Quality of English LanguageN/A
Author Response
Dear reviewer, please refer to the attached file.
Author Response File: Author Response.docx
Reviewer 4 Report
Comments and Suggestions for AuthorsAuthors prepared manuscript titled “Blind edge–retention assessment for PolSAR despeckling methods based on a ratio gradient operator and confidence interval estimation”.
Manuscript is structured unprofessionally. It means: there are three separate parts with different conception and different style of writing. Part 1 – it is introduction and almost all references (twenty eight of them!) we can find in this section. Please notice, that references should be set in Methodology section, too. It is very rare that whole methodology is developed by the authors. Part 2 – It is called “Related work”. This is the section written as a lecture and surly compiled but with two new references only (29,30). References 25– 28 are repeated from the introduction. This is wrong structure. It looks that material is copied. And part 3 – called “Experiments and results”, where we can find the rest, it means data, analysis, processing, results, shortly speaking everything mixed. This is chaotic structure. Please reorganize manuscript.
Detail comments:
Line 58–Should be singular form “it will”
Line 71–Should be “it is well known”
Line 106–What means "Related work" it is unclear? Please do not give a lecture, please provide methodology applied. This part with the current version is not for publication but for teaching students. Please clearly present your objectives and methods. This is your individual research not a public lecture. Please change the lecture form to manuscript form. If, readers want to know principals of polarimetry they will easily find it by themselves. Please describe what you do, how, and using what. With your proposed version the section is long–winded.
Lines 298–311–There are citations repeated from the introduction. This is chaos, still it is wrong structure.
Line 312–Please clarify what the section 4.1 is about?
L.313–Please note that readers are not interested in “many studies”, but they are interested in your particular study. Please change the style of writing. Do not describe "what is happening around the world" but please focus on your particular subject.
Please provide section "Data". Please provide dates for real images.
Lines 379–469–This is very chaotic subsection. Data, processing, results are mixed. What is authors’ rationale for such a order– L,X,C. I suggest doing this way: 4.2.1 PALSAR, 4.2.2. AirSAR, 4.2.3 TerraSAR-x.
Tables 1–4–What means “quantitative assessment values”–please clarify.
Please reorganize manuscript as follows:
Introduction, methodology applied, data, processing, results, discussion (optional), conclusions.
Minor editing of English language required.
Author Response
Dear reviewer, please refer to the attached file.
Author Response File: Author Response.docx
Round 2
Reviewer 3 Report
Comments and Suggestions for AuthorsDear authors,
Thanks for addressing my comments. The revision has improved a lot. I have no more comments.
Author Response
Thanks again for your helpful comments and suggestions, which are quite valuable for us to improve the quality of our paper.
Reviewer 4 Report
Comments and Suggestions for AuthorsThank you for the response letter. I accept authors’ clarifications. However still have doubts about the response 11. If authors’ rationale is to compare single–pol and quad–pol/PolSAR, why authors do not provide particular subsections for that. The idea to use single and quad images is the clou of authors’ research! Authors do not present material for themselves but for the readers. And they (readers) want to follow the concept. Another concept could be –for instance– to compare quad–pol/PolSAR images from different satellites, etc…
Detail comments:
1. Fig. 5–Caption is incorrect. 5(b) Should be: TerraSAR-X image. There is: ALOS image. Please correct.
2. L.416–Please specify image not the figure’s number. It should be "homogeneous SLC".
3. L.416–If authors refer to figure 4 (p.11) after presentation of figure 7 (p.10) it means that the structure of the manuscript is incorrect. Authors must present the figures and the text simultaneously. Please reorder the structure.
4. Section 4.2–Please point out (or/and add subsections) that authors used single–polarization images and quad–pol/PolSAR image. Title of the section “Quantitative Assessment on Three SAR images” is confusing, not informative, and still does not mean anything. In short: chaotic.
5. Table 6. Authors use "ALOS image" but up to page 15, they wrote about “PALSAR image”. Please be consistent, either use ALOS or PALSAR but always the same name for the same data/product. However, I would suggest using ALOS-PALSAR because we have also ALOS AVNIR.
6. Conclusions are chaotic, too. Authors do not summarize their research step by step. But they write that e.g. (L. 526) many studies use optical images to simulate SAR images. But this is new argument you do not use in your research so why authors mention this. That would be again different concept for another article. Please focus on your research and results.
In my opinion authors want to sell/show too much information in one manuscript.
Author Response
Please refer to the attached file.
Author Response File: Author Response.docx
