A Lightweight Swin Transformer-Based Pipeline for Optical Coherence Tomography Image Denoising in Skin Application
Round 1
Reviewer 1 Report
The object of research described in the manuscript is enhancing the quality of OCT skin images using methods based on neural networks.
Both Optical Coherence Tomography and applications of neural networks are area of active research conducted by many group around the world. Therefore, the research results presented in the manuscript are of interest to several research groups.
Novelty of the manuscript is clearly stated in the concluding paragraph of the Introduction.
However, some issues should be addressed prior to the publication of this manuscript.
First, the size of the text in Fig. 2, make it barely legible. Typically minimum recommended size of font is eight points (8 pt). Please modify the figure accordingly.
Second, please revise the manuscript to remove inaccurate language. In particular, ‘submicron-level resolution’ (line 33) typically means resolution below 1 μm, not ~5μm. Other expression, such as ‘has the lowest floating points operation (line 19), ‘scarification of the denoising performance is slight’ (line 369) should also be corrected.
Finally, the language of the manuscript should be improved, as it contains several mistakes. In particular, ‘six repeated’ (line 284), ‘to avoid unstable’ (line 292), ‘Based on the experiment result exhibition’ (line 366), as well as other similar ones should be corrected.
In conclusion, the manuscript cannot be published in its present form, before problems outlined in this review are addressed.
Author Response
Thank you for reviewing our article and giving us comments. Please find our response in the attached file.
Author Response File: Author Response.pdf
Reviewer 2 Report
In this manuscript, the denoising pipeline of a lightweight U-shape Swin (LUSwin)-Transformer is proposed for skin applications. The pipeline can recover high-quality OCT images from the noisy OCT images by utilizing a fast one-repeated OCT scan. The authors show that the proposed method is sufficiently practical with PSNR performance compared with other methods. Therefore, I consider it appropriate to publish this manuscript in its present form.
Author Response
Thank you for reviewing our article and giving us a positive comment.
Reviewer 3 Report
The comments and suggestions are all in the document.
Comments for author File: Comments.pdf
Author Response
Thank you for reviewing our article and giving us comments. Please find our response in the attached file.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
My problems have been well solved and the manuscript has been fully improved.