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.
The paper is well-written, and its structure is easy to follow, with an introduction that explains the motivation behind the work. The relevant references are included throughout. Thank you for your contribution,
[# PeerJ Staff Note - this decision was reviewed and approved by Mehmet Cunkas, a PeerJ Section Editor covering this Section #]
Paper is well-written and structure is easy to follow. Introduction explaining the motivation behind the work. The relevant references are included throughout. All terms, modules, and equations are clearly defined, and the recent formatting and figure clarity improvements are done.
clearly describe the model architecture (HEFFLPNet), its components, and the training process. Included details on datasets, preprocessing, and evaluation metrics( Dice and HD95). The comparison with existing methods is elaborated. The added clarification around random seed usage and statistical significance testing makes the results more credible and reproducible. Including the code or scripts for full reproducibility would be a welcome addition, but overall, the design is rigorous.
evaluated the model by using both quantitative metrics and qualitative visual comparisons and the use of significance testing strengthens these claims. The conclusions are limited to what is supported by the data. Paper discusses its limitations and outlines possible directions for future work.
Please address the remaining issues as requested by the reviewers.
N/A
N/A
N/A
The paper has been significantly revised and polished. The figures and formulas are clear and accurate. The organization and writing are satisfactory. However, the equations in the PDF appear to be in the wrong format, whereas they seem correct in the submitted track-changes Word file.
Thanks for addressing the comments.
One more comment:
- Add more details to Algorithm 1 and refer to the written equations in the manuscript. Reading it like that will not highlight/reflect your contributions that you dedicated time on them.
Minor comments:
- Proofread the manuscript as some typos are found (including spacing and grammatical errors).
- Sort the list of abbreviations A-Z. Add a horizontal line to separate the symbols from the abbreviations.
- I encourage the authors to increase the DPI/resolution of the figures.
- The authors are encouraged to add the AI disclosure statement (For example, Did you use generative AI to assist you in writing this chapter?).
Refer to "Basic reporting" section.
Refer to "Basic reporting" section.
Refer to "Basic reporting" section.
The paper is clearly written and easy to follow, with professional language used throughout. The introduction does a good job of setting up the context and explaining why the study is important. The related work is well-cited and relevant, helping to show how this paper fits into the broader research landscape. The structure of the paper is logical and meets the expected standards, and any technical terms or methods are clearly explained. I did not see any major issues in this.
The article presents a well-structured experimental design that aligns with the journal’s aims and scope. The investigation is rigorous and meets a high technical standard. The authors thoroughly describe the architecture, components (TriCAFE, MSPMA, UpFA), and training process of their proposed HEFFLPNet model. The methodology is clearly articulated, with sufficient details on preprocessing steps, training parameters, datasets (micro-US and CCH-TRUSPS), and evaluation metrics (Dice and HD95) to enable reproducibility. Notably, the authors provide meaningful comparisons with state-of-the-art methods and include comprehensive ablation studies to validate individual module contributions. The code and data links are also shared, supporting transparency. Overall, the experimental design is robust, technically sound, and well-justified. While the methods are solidly implemented, the manuscript would benefit from including additional details on cross-validation strategy, random seed use for reproducibility, and clearer justification for the choice of AG-BCE loss weights (1 and 4).
The experiments are thorough and align well with the stated objectives. Evaluation metrics (Dice and HD95) are appropriate, and the results demonstrate consistent improvements over baseline and state-of-the-art methods. The conclusions are supported by quantitative and qualitative analyses, including ablation studies. The limitations are acknowledged, and future directions are outlined. However, the paper would benefit from including statistical significance tests to validate the reported improvements. Additionally, a deeper discussion of the model’s limitations in clinical deployment scenarios could further strengthen the conclusion.
Please follow the requested mandatory changes strictly.
**PeerJ Staff Note**: Please ensure that all review, editorial, and staff comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
-
-
-
- Insert an introductory paragraph at the beginning of the Related Work section to provide context and summarize the key themes or advancements discussed in the literature.
- The code link appears to be incorrectly formatted. It should read: https://github.com/hse3/HEFFLPNet/tree/main
- Relocate this corrected code link to the Code Availability section under the Declarations part of the manuscript.
- Some sections, such as the Overall Structure section, are written as excessively long paragraphs (e.g., one paragraph spans more than 30 lines). Break these into shorter, more concise paragraphs for improved readability. Review the entire manuscript to address similar issues elsewhere.
- The page containing line 393 (L.393) is empty. Investigate and resolve this formatting issue to ensure content continuity.
- In the Dataset section, properly cite the dataset and include its link to acknowledge the authors or owners who contributed to its creation.
- Enhance Figure 1 by adding detailed annotations or labels to make it more interpretable. For example, some blocks (such as the green ones) currently lack explanatory details or descriptions.
- Organize the list of abbreviations in alphabetical order (A-Z) for clarity and consistency.
- If AI tools were used to assist in drafting the Limitations section, ensure that the content has been thoroughly revised to reflect your voice and expertise. Avoid using em-dashes (a common trait of AI-generated text) and refrain from stating "has several limitations", as this may undermine the significance of your study. Instead, use phrases like "may have limitations". Additionally, refine the "Future Research" paragraph to be more specific and focused rather than overly broad.
- Replace the heading "Analysis of Visual Results" with "Qualitative Analysis" for greater precision and alignment with academic conventions.
- Introduce two new subsections: (1) Privacy and Ethics, which discusses the ethical considerations and privacy safeguards relevant to your study, and (2) Medical Relevance, which highlights the clinical or practical implications of your work.
- Include the overall pseudocode at the end of the Methods section to provide readers with a clear and structured overview of the proposed methodology.
-
-
-
-
-
**PeerJ Staff Note:** Please ensure that all review, editorial, and staff comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
The organization of this paper is clear, and the use of English is acceptable but still can be improved. The introduction and related work provide sufficient context. However, several shortcomings are listed as following:
1) Incorrect literature references. Most of the citations show the error: "Error! Reference source not found.".
2) Redundant content in Lines 94-116. The paragraph beginning with "To address these challenge......" is lengthy and contains redundant information. Much of this content is nearly identical to the description found in the Methods section. Consider shortening or removing it to improve conciseness and avoid repetition.
3) Writing clarity. The writing of this paper can be improved for clarity. Several sentences are unambiguous, which makes them hard to follow and understand.
For example: “Micro-ultrasound segmentation also faces the problem of unclear prostate boundaries, and previous traditional methods lack real-time capabilities due to their focus on 3D volume post-probe sweep. Its resolution is significantly better than that of traditional ultrasound, and its image contrast is also higher, but prostate calcification artifacts are the main challenge.”
4) Several technical and formatting errors. i) inappropriate line break in line 179. ii) "while the deep features of the encoder have a larger receptive field" in line 291, based on my understanding, it seems "encoder" here should be "decoder". iii) Line 295 contains an unexpected space that should be removed.
5) While Figure 2 is professional and visually well-prepared, it is overly complex due to the large number of lines and components. The mechanism it presents is difficult to interpret, especially without a step-by-step breakdown or annotated explanation. Consider simplifying the figure or providing a more detailed description in the caption or main text.
Article content is within Aims and Scope of the journal and article type. Sufficient details are given for replication, including code and dataset link, data preprocessing, hyperparameter settings. Plus, evaluation methods, assessment metrics are adequately described.
The motivation and rationale of this work are clearly stated. Limitations and conclusions are included and well stated. The experimental setup is generally sound and the results are presented in a clear manner. However, several aspects of the paper require improvement before it can meet the standards for publication:
1) While the results are clearly presented and appear sufficient, the methodology section lacks the necessary detail and clarity. The current description is overly general and vague, making it difficult for readers to fully understand the proposed approach and replicate the work.
2) The performance improvement over MicroSegNet is marginal, especially considering the additional parameters introduced by the proposed modules. The authors should provide further justification for the architecture designs.
2) The ablation study is not considered exciting, since the observed performance difference between configurations is only about 0.1%, which raises concerns about the actual effectiveness of the proposed components. More in-depth analysis or additional experiments are needed to validate the effectiveness of each module.
Abstract: The abstract is overly lengthy and should be revised for conciseness to better reflect the core contributions and findings of the work.
Citations: Several references are incorrectly cited or missing, resulting in "Error! Reference source not found." Please ensure all citations are properly formatted and cross-referenced throughout the manuscript.
Equation 1: The presentation of Equation 1 is unclear. It should be broken down into more concise components and explained in a more structured and demonstrative manner.
Abbreviations and Symbols: The authors should include a dedicated table listing all abbreviations and symbols used in the paper for clarity and ease of reading.
Symbol Formatting: The formatting of mathematical symbols should be standardized throughout the manuscript to improve readability and consistency.
Figure 1: Additional details should be added to Figure 1 to enhance its explanatory value and support better understanding of the methodology or results.
Figure 2: The lines in Figure 2 should be colorized to improve visual clarity, as the current grayscale version makes it difficult to distinguish between different elements.
Failure Cases: The authors should include additional examples where the proposed system failed to perform accurate segmentation, along with a discussion of potential reasons for such failures.
Additional Experiments: Two new experiments should be included:
- One evaluating the system's performance on images/datasets containing multiple objects.
- Another assessing performance on images containing small-scale objects, such as tumor regions or blood vessels.
Check "Basic reporting" section.
Check "Basic reporting" section.
Check "Basic reporting" section.
The Introduction provides a clear overview of the problem domain (prostate segmentation in micro-ultrasound imaging), motivation, and the significance of the work. It does a reasonable job of explaining the limitations of current methods (e.g., MicroSegNet and TransUNet) and establishes the rationale for HEFFLPNet.
However, the introduction relies heavily on unsupported claims (e.g., "Micro-US offers higher resolution...") without early, accessible citations. Additionally, it lacks an explicitly stated research gap in the final paragraph before introducing the proposed solution.
Strengthen the introduction by integrating proper citations earlier and clearly articulating the specific gap this study addresses.
The discussion of preprocessing is adequate, but the rationale behind some decisions (e.g., why 224×224 resolution was chosen, or why a pixel intensity range of [0,1] was used) could be briefly justified.
The manuscript suffers from numerous broken citations (e.g., “Error! Reference source not found.”). This severely undermines the credibility of the related work and methodology justification.
The experiments and evaluations are performed satisfactorily, including ablation studies that isolate the contributions of individual components (TriCAFE, MSPMA, UpFA). These ablation results are informative and help substantiate the effectiveness of the proposed modules. No Comment
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.