Blood Cell Revolution: Unveiling 11 Distinct Types with ‘Naturalize’ Augmentation

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript explores state-of-the-art deep learning techniques, including pre-trained ConvNets, 4 ViTb16 models, and custom CNN architectures for PBC classification. The experiments are implemented to testify to the feasibility and performance of the algorithms along with a novel “Naturalize” augmentation algorithm. The text is clearly written. The authors describe in detail and with interesting scientific details the experiment performed.

The following remarks should be considered:

a.      In Fig. 6 the newly generated composite PMY image contains more than 9 cells, even if the last step refers to 1 BCK, 1 WBC, and 4-8 RBC images. Please explain.

b.      The introduction should have at least one phrase regarding the “Naturalize” augmentation method, otherwise, it will make the reader jump directly to the 3.3 subsections.

c.       In subsection 3.4, you compared „Naturalize” and conventional augmentation techniques. Moreover, conventional data augmentation is in the already augmented data preprocessing step 2. How does this affect the classification? Please explain.  

 

 

All the best,

Author Response

"Please see the attachment."

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript presents an approach in blood cell classification, expanding the five distinct blood cell types to 11-class classification system. I would request the author to address the following issues:

1) It is recommended that the abstract begins with a concise motivation including the context, before detailing the accomplishments of the work.

2) Did the authors consult with medical experts to validate the feasibility of expanding the blood cell classification in their study, and if so, how did this contribute to the study's credibility?

3) Please include a table that compares performance metrics, specifically accuracy, with results obtained by other researchers who have utilized the same dataset.

4) In Line 44, clarification is needed for "intricacy of granulopoiesis 1." An explanation or reference is required.

5) The manuscript should explicitly state the software or system used for Exploratory Data Analysis (EDA).

6) Why do accuracy, macro avg, and weighted avg appear to be the same for the Support column in Table 6-10? It seems unclear how this column aligns with the rest of the data.  Could you provide clarification on this?

7) A more comprehensive discussion on the limitations and challenges encountered during the experimentation process is required.

Comments on the Quality of English Language

Minor editing of English language required

Author Response

"Please see the attachment." 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have addressed most of my comments in their revised manuscript. I have the following observations in the revised manuscript:

1) I would like to draw your attention to a consistency issue in the formatting of Table 5- 12. It appears that there is a variation in the decimal point usage for precision, recall, F1-score, accuracy, etc., across these tables. For instance, Table 5 displays values in percentage format with two decimal places, while Table 6 exhibits the full values with no decimals. Furthermore, Tables 7-12 present values without using the percentage format. I recommend maintaining a consistent format for these metrics throughout Tables 5-12.

2) I still don't understand the Support column's result in Table 7-11. For instance, how the accuracy, macro avg and weighed avg is same 2200 in the Support column of Table 11?

Comments on the Quality of English Language

Minor editing of English language required

Author Response

 "Please see the attachment."

Author Response File: Author Response.pdf