MT-VTON: Multilevel Transformation-Based Virtual Try-On for Enhancing Realism of Clothing
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors1.Please unify the terminology about clothing, use clothing or garment.
2.Please use the original dataset name in [*]: Zalando Dataset. Note that, this dataset is not publicly available, if you want to exhibit their images in your manuscript, please get the agreement of the first author of [*].
[*] VITON: An Image-based Virtual Try-on Network
3.About abbreviation, please follow the abbreviation behind its full name for reading convenience, such as Thin-Plate Spline (TPS)
4.More TPS-based virtual try-on methods need to be discussed in related works. For example,
[*] Toward multi-category garments virtual try-on method by coarse to fine TPS deformation
[*] Toward Realistic Virtual Try-on Through Landmark Guided Shape Matching
[*] PG-VTON: A Novel Image-Based Virtual Try-On Method via Progressive Inference Paradigm
5.Please describe the prior (segmentation and pose) processing methods, and other details like the category in segmentation also are needed.
6.About Eq.4, I know l^ and l mean predicted result and the ground truth, but please give their description.
7.Since you set a lot of hyper-parameters in Eq.6, please list their specific values in Sec.5.1.
8.Some sentences have grammatical mistakes, please check all your manuscripts. Some examples are enumerated as:
‘our approach surpasses other baseline methods in terms of quantitative performance’
‘While all images 258 have dimensions of 1280 × 720, we resize them to dimensions of 512 × 256 after cutting out 259 white spaces in the images.’
9.Please consider discussing some limitations of your work and pointing out some future research directions.
Comments on the Quality of English LanguageSome sentences have grammatical mistakes, please check all your manuscripts.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThank you for the opportunity to review the article “MT-VTON: Multilevel Transformation-Based Virtual Try-On For Enhancing Realism”. The paper addresses an interesting and well researched theme in the recent period about the implication of using and applying Virtual 3D fashion solutions for the consumers to interact with fashion items online.
This study represents a solid effort in the field approached. It is constructed in a mature manner, following the publication standards of the journal, discussing the subject that needs to be comprehensively analyzed.
Also, the study is written in an adequate manner, with a specific review of the literature and robust research design. The results are presented clearly and coherently, using visuals and text. The images and tables presented in the paper were relevant to explore the results of the research and the ways these were adapted to the explanations in the text.
Moreover, there are some point-by-point observations that should be addressed in this revision.
- The title should include the concept of cloth/ clothing or fashion or fashion items online.
- Also, these concepts should be present as keywords.
- Resolve text issues: Line 56> experiments. comparing, Line 335 In Proceedings of the Proceedings of… (this problem appears several times in References section).
- Conclusion section should be better presented and developed.
Comments on the Quality of English Languagenot the case
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe novelty of the proposed paper is reflected by the key contributions associated with the multi-faceted transformation strategy that operates at both pixel and image patch levels. This method enables a more accurately complex deformations of clothing and preserve intricate textures.
The research paper is well-organized and follows the template of the Applied Science Journal. The introduction provides a good overview regarding virtual clothes try-on and there are a good number of cited references.
The structure of the paper along with the research papers are presented at the end of the introduction section.
The related work section integrates a good number of recently published papers. This section is divided into three main categories, and it provides a good overview of each category of methods, as the advantages and disadvantages of each category are being discussed.
The proposed approach involves five sub-processes which are illustrated within Figure 1. The workflow of the proposed method is well defined starting with the cloth segmentation, segmentation warping, cloth warping, cloth refinement and the final try-on process that illustrates the final output.
Within the next section, the paper presents the details of each of the five sub-processes involved in the proposed approach to cloth refinement for virtual try-on.
The following section is well detailed indicating the datasets. The proposed method was trained on the Korean fashion dataset maintained by AIHub for which the authors have also provided a link. This dataset contains images of flattened clothes, which are more closely related to real-world conditions than the images in the VITON dataset which is available on GitHub.
The proposed paper provides both qualitative and quantitative results to demonstrate the effectiveness of the proposed approach which is compared to existing approaches.
The paper is clear, concise, and well-written. I recommend that the paper be accepted for publication in the current form.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsThank you for the opportunity to review the revised version of the paper “MT-VTON: Multilevel Transformation-Based Virtual Try-On For Enhancing Realism”.
The authors responded and explained with reasonable arguments and corrected all the remarks and observations highlighted in the previous review and the results suggest a more consistent and logical text.
To sum it up, the authors developed a more in-depth theoretical presentation about the subject, integrating the suggested aspects of the review.
I consider that the paper is publishable after a final check from the authors.
Comments on the Quality of English Languagenot the case
