Next Article in Journal
Automated Detection of Left Bundle Branch Block from ECG Signal Utilizing the Maximal Overlap Discrete Wavelet Transform with ANFIS
Next Article in Special Issue
Building DeFi Applications Using Cross-Blockchain Interaction on the Wish Swap Platform
Previous Article in Journal
Can We Trust Edge Computing Simulations? An Experimental Assessment
Previous Article in Special Issue
Traffic Request Generation through a Variational Auto Encoder Approach
 
 
Article
Peer-Review Record

Functional Data Analysis for Imaging Mean Function Estimation: Computing Times and Parameter Selection

by Juan A. Arias-López 1,2,*, Carmen Cadarso-Suárez 1,2 and Pablo Aguiar-Fernández 3,4
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Submission received: 6 April 2022 / Revised: 19 May 2022 / Accepted: 25 May 2022 / Published: 2 June 2022
(This article belongs to the Special Issue Selected Papers from ICCSA 2021)

Round 1

Reviewer 1 Report

Although not my specific field, I found that overall there seems to be a lack of detail of what is being calculated. While the references may provide greater insight, readers do not generally have the time to familiarize themselves with all of the details. It is important to provide an overview of the methods to a level of detail enough that the reader can understand the basics of what is happening. Here I found that I could not really understand how the bicubic spines were calculates and combined across the various datasets.

It is very difficult to understand what datasets are being compared and how, of if the basic goal of the paper was to compare the efficiency of Delaunay triangulations as a surrogate for the actual analysis. If that is the case, using R to perform this task may not be efficient compared to a dedicated compiled program.

Most readers of the journal will be unfamiliar with FDA, so it will be necessary to explain the steps in detail.

A few specific comments appear below.

Line

Comment

55

Leakage appears to be a big problem but is not defined. It is not a standard term for image analysis

58

Replace “Problematic” with Problem

118

Replace “drawn” with draw

118-128

It appears from this description that images from 110 patients were somehow coaligned and registered, but the description needs to be included in more detail. Were the normal and AD patients handled separately? What role did the MR play and where did it come from – is this part of the ADNI data you used? What is “unwrapping”?

It is unclear if all images are being independently being handles or somehow all are being combined together. This section needs to be expanded and rewritten.

130

The Delaunay triangulations shown in Fig 1 appears to be applied to a single slice of a segmented CT or MR of a single brain, but it is not clear how this relates to the PET data from line 119. Was this used just as a test for computational cost assessment?

81

The authors are proposing representing the images using bivariate splines, but this seems unclear how this will be done. Will the grey scale of each slice be used as the z dimension of the spline, with x and y representing the grid coordinates? Of so this needs to be stated clearly.

144

What is alpha?

140

“the estimation of a group of images’ mean function and its associated SCC in the form of images”. Very confusing. Assuming a bivariate spline has been calculated for a single slice, it is not clear how a group image can be coalesced from disparate datasets of differing dimension, scale, orientation etc. unless the images are somehow mapped to an atlas or some kind of normalization has occurred.

149

See the comments of line 140 above. There is no clarity as to how the comparison is being made.

Fig 3

What are the two samples? Are they different patients, different groups etc.

156

While the effect of the number of points used for the Delaunay triangulations seems to be the main

38

It’s unclear exactly what the functions are representing. The authors indicate that the functions vary in time, but the PET images are generally static.

82

We need to know the basics of how to calculate the SCC

 

The paper spends a lot of time determining the computational cost of implementing the FDA algorithm however only CPU based calculations were used. Is it appropriate to consider using GPU based acceleration for this algorithm? Also R is an interpreted language and can be slow. Some comments on this aspect might be useful.

 

Author Response

Dear reviewer, find attached my response to your constructive criticism and suggestions. Thanks for this opportunity.

Author Response File: Author Response.pdf

Reviewer 2 Report

Please see attached.

Comments for author File: Comments.pdf

Author Response

Dear reviewer, thank you for your overall criticism of my article. Find attached a document with my responses and changes I have made accordingly. Again, thank you for the opportunity to improve my work.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript presents a research on computing times and parameter selection for Functional Data Analysis (FDA) applied to imaging mean function estimation.
The obtained results confirmed past research finding and confirms that this method should be further studied and applied to the field of medical imaging.
I find the topic interesting and being worth of investigation and the document is well strucutred, organized, fluidly written, has enough background information, the methodology followed is clearly explained, formulas are correct, the results are clearly presented.
Although I propose the following comments/suggestions:
- The abstract is poorly descriptive of the content, it should be better organized: problem, motivation, aim, methodology, main results, further impact of those results.
- keywords are inexistent should be in alphabetical order.
- I strongly suggest authors from refraining using personal pronouns such as "we" and "our" throughout the text and I encourage them to write it in an impersonal form of writing.

Author Response

Dear reviewer, thank you for the opportunity to improve my research with help of your suggestions and constructive criticism. Find attached a document with details on the changes I performed and answers to your suggestions. Thank you.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have addressed most of my concerns. Thank you. A few minor editorial issues remain.

 

Author Response

Thank you, Reviewer 1, for the feedback provided in order to improve my research. I am glad to hear from you after submitting my revisions and happy to know that you now consider that your concerns on my article are now addressed. 

Back to TopTop