Next Article in Journal
Investigation of UAV Detection by Different Solid-State Marine Radars
Previous Article in Journal
Service Discovery Method Based on Knowledge Graph and Word2vec
 
 
Article
Peer-Review Record

A Segmentation Algorithm of Colonoscopy Images Based on Multi-Scale Feature Fusion

Electronics 2022, 11(16), 2501; https://doi.org/10.3390/electronics11162501
by Jing Yu 1,2, Zhengping Li 1,2,*, Chao Xu 1,2 and Bo Feng 1,2
Reviewer 1:
Reviewer 2: Anonymous
Electronics 2022, 11(16), 2501; https://doi.org/10.3390/electronics11162501
Submission received: 13 July 2022 / Revised: 4 August 2022 / Accepted: 9 August 2022 / Published: 11 August 2022

Round 1

Reviewer 1 Report

This study proposes a segmentation algorithm for colonoscopy images based on multi-scale feature fusion. The work looks interesting. My assessments are listed below.

1. In medical image segmentation, metrics that depend on TP, FP and FN values such as Dice, F1, Recall are used for performance measurement. Calculation of TN is usually not required because too many bounding boxes are proposed. Why did the authors calculate the TN value in this study and why did they use the accuracy metric in a segmentation problem?

2. Figure 4 is not clear. Which segmentation outputs belong to which method has to be presented more clearly and with more images.

3. Figure 5 shows the graphical representation of the values in Table 2, that is, the same results are presented. Instead, more meaningful box plots can be used.

4.The authors' evaluation of the results of their own study and the results of previously proposed studies with statistical significance tests will increase the scope and quality of the study.

5. The segmentation scores of the proposed method in the study can also be compared with Mask R-CNN.

6. More up to date references have to be included in the study. There are no references from 2022 and only 3 references from 2021.

7.The change of the loss values of the proposed network during the epoch can be added.

8. Why did the authors calculate the Dice metric with a confusion matrix and not based on cluster similarity (join and intersection) operations? How will the results change if Dice is calculated with cluster similarity?

Author Response

Dear Reviewers,

Thank you very much for your time involved in reviewing the manuscript We also appreciate your clear and detailed feedback and hope that the explanation has fully addressed all of your concerns. attachment is my point-by-point response to your comments, Please see the attachment.

 

With Kind Regards

Prof. Zhengping Li.

Author Response File: Author Response.pdf

Reviewer 2 Report

An interesting research contribution  "A segmentation algorithm of colonoscopy images based on multi-scale feature fusion" from the authors. Few queries and suggestions are as below:

(1) Feature extraction is a prerequisite for image segmentation. Here authors use the pyramid feature extraction module. As you said of proposing a segmentation algorithm than name your method if you feel that your method is novel. 

2) Good results shown on "Multi-proportion Fusion Module" and comparison of the Index results of different algorithms.

(3) Can we apply same approaches with other cancer type images? Include in future directions of your research contribution.

(4) Authors may write about the  significance of the contribution and future direction.

(5) Include other latest references related to multi-scale feature fusion.

Author Response

Dear Reviewers,

Thank you very much for your time involved in reviewing the manuscript We also appreciate your clear and detailed feedback and hope that the explanation has fully addressed all of your concerns. attachment is my point-by-point response to your comments, Please see the attachment.

 

With Kind Regards

Prof. Zhengping Li.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have corrected the revisions I have indicated in the previous version of the study. 

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.


Back to TopTop