Validated Filter-Based Photoreceptor Count Algorithm on Retinal Heidelberg High Magnification Module™ Images in Healthy and Pathological Conditions

Round 1

Reviewer 1 Report

This paper discussed cone count method for OCT and the results were verified by comparing graders. Followings are my opinion.

1. Commercial names were too frequent in the text, compared with other references. The commercial names should be decreased and substituted into other general words if possible. For example, HMM was just used as High Magnification Module in other papers. You can check the issues of OCT in this journal for the exposure level of the commercial names.
2. Image quality is poor. The image resolution is insufficient for journal publication and must be increased. Fig. 1 and 3 are the most important, but the images are too blurred.

Author Response

1. Commercial names were too frequent in the text, compared with other references. The commercial names should be decreased and substituted into other general words if possible. For example, HMM was just used as High Magnification Module in other papers. You can check the issues of OCT in this journal for the exposure level of the commercial names.
   As suggested by the reviewer we have replaced the commercial name ‘HMM’ with the general description ‘high magnification module’ to reduce the frequency of commercial names.
   
   
2. Image quality is poor. The image resolution is insufficient for journal publication and must be increased. Fig. 1 and 3 are the most important, but the images are too blurred.
   We converted our .jpeg images into .tiff files and increased the resolution of all figures to 900 dpi to meet journal requirements. Additionally, we requested Ms. Aria Tang (Assistant Editor) to review the quality of the new figures.
   
   

Reviewer 2 Report

*) If the authors deem it appropriate, it would be more useful to collect the types of patients in a single table.

*) I think it is appropriate to insert some mathematical details regarding the filter used.

*) The presence of acronyms in the text would require the insertion of a table that lists them all in order to increase the readability of the text.

*) Please explain in a little more detail how to obtain the images.

*) The images obtained are interesting and, even more, their processing. The results obtained are certainly valuable. However, I believe that such images could be affected by uncertainties so, strictly speaking, fuzzy preprocessing techniques should be applied. Obviously, such an approach would require an overall effort that goes beyond the present work. However, I advise authors to insert a sentence in the text that highlights this possibility by putting the following relevant works in the bibliography:

doi: 10.1109/ACCESS.2017.2776349

doi: 10.1016/j.asoc.2020.106452

Author Response

*) If the authors deem it appropriate, it would be more useful to collect the types of patients in a single table.
Only one specific type of patient was investigated in our study, namely patients with a p.Arg142Trp mutation in the PRPH2 gene. This genotype resulted in a CACD phenotype in all of our patients.(line 63-64 and 172-174) To facilitate future high magnification module research we have already provided a supplementary table listing the cone count per regions of interest (ROI) of all study participants. In this table relevant characteristics (such as age and refractive error) of each study participant are also listed. To clarify the type of participant in the table we used pseudonyms beginning with ‘Healthy’ for all healthy subjects and ‘CACD’ for all patients.  

*) I think it is appropriate to insert some mathematical details regarding the filter used.
We agree with the reviewer and inserted mathematical details regarding the utilized high-pass filter (line 133) and Gaussian function (line 157).

*) The presence of acronyms in the text would require the insertion of a table that lists them all in order to increase the readability of the text.
We have inserted a table prior to the introduction listing acronyms in alphabetical order to increase readability of the text.

*) Please explain in a little more detail how to obtain the images.
To increase reproducibility of our image acquisition protocol we have added more detail how we obtained high magnification module images (line 80-100).

*) The images obtained are interesting and, even more, their processing. The results obtained are certainly valuable. However, I believe that such images could be affected by uncertainties so, strictly speaking, fuzzy preprocessing techniques should be applied. Obviously, such an approach would require an overall effort that goes beyond the present work. However, I advise authors to insert a sentence in the text that highlights this possibility by putting the following relevant works in the bibliography:

doi: 10.1109/ACCESS.2017.2776349

doi: 10.1016/j.asoc.2020.106452
Excellent suggestion by the reviewer. We have added a sentence in the discussion highlighting the possibility of fuzzy preprocessing techniques (line 287-289) and added the suggested literature in the bibliography (line 422-425).