Automated Scattering Media Estimation in Peplography Using SVD and DCT

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The topic of the paper is innovative, proposing the idea of automating the estimation of scattering medium information through SVD and DCT methods, and improving the existing Peplography method. The research work is related to practical needs, and experimental results show that the method can achieve better results in different scenarios. However, there are still some logical, detailed, and structural issues in the paper that need further revision and improvement. The following are specific review comments.

 

1. The description of the research background needs to be more detailed. The introduction section lacks sufficient discussion on the challenges of existing technologies and the advantages of Peplography. Comparative analysis can be added to strengthen the necessity and background depth of the research. Suggest supplementing the requirements and impacts of automated scattering medium estimation in practical application scenarios to enhance persuasiveness.

2. The algorithm description requires a more detailed explanation. The description of the combination process of SVD and DCT methods is relatively vague, and the physical meanings of some formulas (such as formulas 16 and 17) are insufficiently explained, which needs further clarification. Suggest adding algorithmic pseudocode or more specific flowcharts for readers to have a more intuitive understanding of the method.

3. The experimental part needs to strengthen comparison and analysis. The current experiment only presents a simple comparison with traditional methods, lacking a more comprehensive performance evaluation. More experimental results of benchmark algorithms can be supplemented, and the advantages and disadvantages of different algorithms can be analyzed in detail.

4. Transparency of datasets and experimental settings. The scenario settings and data generation process used for testing were not described in detail. Suggest disclosing more experimental details and data sources to improve the reproducibility of experimental results.

5. Lack of discussion on limitations. The possible limitations of automation methods in complex environments, such as when multiple light scattering sources are present simultaneously, have not been discussed. Suggest adding an analysis of the limitations of the method under different conditions and potential improvement directions.

6. The literature review section needs to be expanded. The current literature review is insufficient to cover the main research advances in the current field, especially the latest developments combined with deep learning techniques. Suggest supplementing important literature in the fields of scattering medium processing and 3D imaging in recent years.

7. Expressing issues and revising language. There are multiple grammar errors and unclear expressions in the paper, which require comprehensive language polishing. Expressions such as' can generate the distortion of the image 'are too colloquial, it is suggested to use academic language instead.

8. Chart improvement. The quality of charts (such as Figure 6 and Figure 7) needs to be improved. It is recommended to unify the style and ensure high resolution. Some charts do not indicate the meaning of units or parameters, and additional explanations should be provided.

9. The conclusion section needs to be more specific. The current conclusion section is too brief, and a summary of the contribution of the method should be added, along with an outlook on future work. For example, the future direction of combining AI models is only briefly mentioned and can be further explained in detail.

10. Title optimization. The current title is lengthy and lacks academic rigor. It is suggested to optimize it to a more concise title that highlights key technological innovations, such as "Automated Scattering Media Estimation in Pepperography using SVD and DCT".

11. To support further research and applications in remote sensing multi-object tracking, I recommend making the code available through a dedicated website.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This paper presents a novel method for automating scattering media estimation in peplography using Singular Value Decomposition (SVD) and Discrete Cosine Transform (DCT). The research addresses a critical limitation of conventional peplography by automating the estimation process, reducing manual intervention, and improving efficiency in real-world applications. The proposed approach is validated through comprehensive experiments and image quality assessments, demonstrating its effectiveness across different conditions.

 

1. Inadequate Methodological Explanation:

• The methodology section introduces SVD and DCT but lacks a cohesive link between the two techniques. The reasoning behind applying SVD after DCT is not sufficiently justified. Furthermore, the role of Gaussian low-pass filtering is underexplored, leading to potential confusion about the contribution of each step.

    Action Required:

• Clearly articulate the theoretical basis for combining DCT and SVD. Explain why truncating singular values is superior to conventional filtering techniques for scattering media estimation. Provide explicit mathematical transitions between DCT, SVD, and Gaussian filtering.

 

2. Experimental Setup:

• The experimental setup is well-described, but additional details on the specific environmental conditions (e.g., fog density levels, light intensity) could improve reproducibility.

• The choice of the 5x5 camera array requires justification—why was this array configuration selected, and how does it generalize to larger or smaller arrays?

 

3. Results Presentation and Interpretation:

• The visual comparisons (Figures 10-11) lack clarity. The images are small, and critical details are difficult to discern. No zoomed-in views or edge-detection comparisons are provided, which hinders thorough evaluation.

• The IQA (Image Quality Assessment) metrics are presented without sufficient interpretation. While the numerical results suggest performance parity, the implications of minor deviations in SSIM or FSIMc are not discussed.

 

4. Lack of Discussion on Limitations:

• The paper lacks a critical evaluation of its own limitations. Issues such as computational cost, scalability to real-time applications, and potential artifacts from Gaussian low-pass filtering are ignored.

Action Required:

• Include a limitations section discussing computational overhead, potential failure modes, and scenarios where the proposed method might underperform. Address the scalability of the method to video streams or large-scale datasets.

5. Future Work Lacks Technical Depth:

• The future work section mentions the integration of AI but provides no concrete pathway for achieving this. The vague mention of AI integration diminishes the paper’s forward-looking value.

Action Required:

• Specify the types of AI models (e.g., transformer networks, self-supervised learning) and explain how they could enhance scattering media estimation. Propose a preliminary workflow or outline future experimental directions in detail.

 

 

Comments on the Quality of English Language

The quality of English in the paper requires significant revision to improve clarity, readability, and professionalism. Key issues include grammatical errors, awkward phrasing, and inconsistent terminology. These language issues may obscure the paper’s contributions and reduce its overall impact.

 

Example Revision (Abstract Section):

Some sentences are unnecessarily complex or awkwardly structured.

Example: “In the conventional scattering media removal methods, it can remove the light scattering media in the scene by utilizing various image processing techniques…”

Suggested Revision: “Conventional scattering media removal methods reduce light scattering in images using various image processing techniques…”

 

Typographical Errors:

“Discrete cosine transfrom” → “Discrete cosine transform”

 

Repetitive Language:

Certain phrases are repeated excessively, affecting readability.

“The scattering media information can be estimated by truncating the singular value matrix…” appears multiple times with minimal variation.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This paper is written logically and academically.  The manuscript has a solid theoretical background, and the references are used properly for the addressed subject. The authors bring a valuable contribution to the existing literature on peplography by suggesting a method that is automating the estimation of media information that is scattered. 

 

The methodology section and mathematical context are rigorously explained and described. The authors talk about the used image quality assessment mertics such as SSIM, FSIMc, GMSD, LPIPS. The authors take an innovative approach by utilizing SVD and DTC. This also highlights the deficiencies of classic peplography techniques. The used flowchart enhances the understanding of the complexity of the mathematical framework. 

 

Please find below my recommendations that could enhance the impact of this paper:

 

The concept “bulk edge” could benefit from further explanation and elaboration in the singular matrix.

 

The experimental results are presented. However, I suggest a more detailed description of the limitations. 

 

I also suggest a comparison with similar top-notch automated scattering media removal techniques. 

 

Finally, a discussion about elevating and optimizing classic peplography could be a great addition to the paper.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Most of the modifications have been completed as required. It is recommended to improve the clarity of the images and refine some of the language.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have provided a comprehensive response to my  comments, addressing the identified issues thoroughly. Their revisions include:

1. Improved Methodological Clarity:

   • Enhanced explanations of the rationale for combining SVD and DCT and their roles in the proposed method.
   • Detailed experimental setups, including environmental conditions and justifications for the 5x5 camera array configuration.

2. Results and Limitations:

   • Revised figures for clarity, including high-resolution images and detailed comparisons.
   • Added a limitations section, discussing computational costs, scalability issues, and potential artifacts, as well as future plans to address these challenges using AI models.

3. Language and Readability:

   • The language has been improved significantly, addressing grammatical issues, awkward phrasing, and repetitive language.
   • Key sections, such as the abstract and introduction, now better highlight the contributions and innovations of the study.

4. Future Work:

   • The future directions are now more specific, including potential AI integration methods and their expected impact.