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 authors have reformatted the referencing style as per the reviewer's recommendation. And updated the word 'ours' as well. I will be happy to recommend it for acceptance following the reviewers' comments.
[# PeerJ Staff Note - this decision was reviewed and approved by Xiangjie Kong, a PeerJ Section Editor covering this Section #]
The reviewer has agreed that his/her comments are addressed. There are minor comments that the authors need to address. Other than those, the paper is now well shaped and can be accepted. Hence, I am recommending a minor revision.
This has been improved considerably.
The experimental design is well elaborated now.
Findings have been validated and compared results from other authors.
1. There is the use of "ours" in some places in the paper; the name of the model is an appropriate idea, as accepted in the comments.
2. Please revisit your pattern of referencing other's work. I am not sure whether "Haines et al.(Haines et al., 2016) introduced" is a good idea to refer to the work or not. Please look into this.
The reviewers have identified that the paper needs careful revision for English composition.
Overall, the reviewers agree that the paper contributes to the topic of handwriting font generation, and the need is well elaborated i.e. to achieve better feature extraction.
Thus the paper is recommended for a second round of submission after careful addressing of the comments provided by the two reviewers.
[# PeerJ Staff Note: The review process has identified that the English language should be improved. PeerJ can provide language editing services - please contact us at [email protected] for pricing (be sure to provide your manuscript number and title) #]
Clear and Unambiguous, Professional English Used Throughout:
The manuscript is generally well-written, but it suffers from minor grammatical errors and awkward phrasings that could impede understanding for an international audience. For instance, the sentence structure in the abstract and introduction could be simplified for clarity.
Suggested Improvements:
I recommend a thorough proofreading by a fluent English speaker or a professional editing service to enhance readability and ensure the text is clear and unambiguous.
Literature References, Sufficient Field Background/Context Provided:
The manuscript does a commendable job of situating its contribution within the existing body of work on handwriting font generation. It references relevant prior literature, demonstrating an understanding of the field and its challenges.
Suggested Improvements:
No significant improvements needed in this area. The coverage of relevant literature and the contextualization of the research are adequate.
Professional Article Structure, Figures, Tables. Raw Data Shared:
The structure adheres to standard scientific reporting formats, with clear sections for the introduction, methods, results, and discussion. However, the integration of figures and tables into the narrative could be improved for clarity.
Suggested Improvements:
Ensure that each figure and table is discussed explicitly within the text, highlighting their relevance and how they support the manuscript's arguments.
Original Primary Research within Aims and Scope of the Journal:
The manuscript presents original research that fits well within the journal's scope. The PSA-HWT model is a novel contribution to the field of handwriting font generation.
Research Question Well Defined, Relevant & Meaningful:
The research question is clearly defined and addresses a known gap in the field—the need for improved feature extraction in handwriting font generation models.
Suggested Improvements:
The manuscript could benefit from a more detailed explanation of the choice of the pyramid squeeze attention mechanism over other potential solutions.
Rigorous Investigation Performed to a High Technical & Ethical Standard:
The investigation appears to be rigorous and technically sound. The methods section provides sufficient detail for reproducibility.
Suggested Improvements:
A discussion on any ethical considerations related to the research would enhance this section, even if to confirm that ethical issues are not directly pertinent to this study.
All Underlying Data Have Been Provided; They Are Robust, Statistically Sound, & Controlled:
The manuscript utilizes the IAM dataset for evaluation and presents a detailed analysis of the model's performance using appropriate metrics. However, a clearer statement on the availability of raw data and code would strengthen the manuscript.
Suggested Improvements:
Explicitly state the availability of all research artifacts necessary for replication, including different raw datasets/scripts, and detailed configuration settings used during experiments.
Conclusions Are Well Stated, Linked to Original Research Question & Limited to Supporting Results:
The conclusions are appropriately stated and well-supported by the results. The manuscript successfully links its findings back to the original research question.
Suggested Improvements:
Further discussion on the implications of these findings for future research and potential applications would be beneficial.
The manuscript "PSA-HWT: Pyramid Squeeze Attention-Handwriting Transformer for Handwritten Font Generation" makes a significant contribution to the field of handwriting font generation. With minor revisions for clarity, detail, and accessibility, particularly regarding the integration of figures and tables, raw data availability, and minor language polishing, this work would make a valuable addition to the literature.
The paper “Pyramid squeeze attention-handwriting transformer for handwritten font generation” must be improved and revised before resubmission.
(1). The paper needs to be checked for English and grammatical mistakes.
(2). It is advised to use Latex for manuscript writing.
(3). The introduction and literature review section needs to be further improved.
(4). The literature also needs to be presented in tabular form.
(1).For readers’ point of view, It would be beneficial if Authors include some more text about existing studies in the related work section.
(2). It should be explained in detail how the present study results/methodology is better.
(3). Comparison of results needs to be more explained.
(1). There is need to give algorithmic view of proposed methodology.
(2). For readers clarity, it will be good to explain proposed method graphically as well as using example cases.
(3). Failure cases analysis and general observations of proposed method are completely missing in the study.
.
This paper has been worked on well. The authors have used a good number of references, and these are related ones.
They have worked on a publicly available dataset and partitioned this into train and test sets. This is a good effort to improve the effectiveness of a process with the help of related developments.
The authors have put a few pieces together to improve the process of handwritten font generation.
The research question is well-defined, and the authors have done a good number of experiments to come to their conclusions.
No comments.
This well-written paper addresses a significant problem of handwritten font generation by machines. I find a paradox here: handwritten but generated by a transformer. Authors may like to find a better idea to present the situation.
I have the following suggestions to incorporate before taking a final call on this paper.
Line 48: "(GANs) (" should be used in place of "(GANs)(". A similar modification should also be done in other similar situations.
Use of white space is very confusing at many places, please look into this. I mean, we should have a space after the comma and also after a full stop.
Line 113 to 134: The use of symbols will confuse the potential readers. I am sure the authors do not mean here that K and italicized K are two different symbols in their work.
I am not sure how replacing a 7×7 convolution kernel in the ResNet50 network with three 3×3 convolution kernels will increase the model's expressive capabilities. We are reducing the number of parameters here.
The grammar needs to be revisited, for example, the use of verbs in Line 257.
In Tables 1, 2 and 3, "ours" should be replaced by the name of the model, I believe. Proposed Model (PSA-HWT) can be a better choice if the authors feel so.
In the Line 275, "highest" is wrongly used.
Please look into the paragraph in the lines 289 to 296. This needs to be revisited.
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.