Next Article in Journal
Radiation Hardness Study of Single-Photon Avalanche Diode for Space and High Energy Physics Applications
Next Article in Special Issue
Automatic Personality Assessment through Movement Analysis
Previous Article in Journal
Experimental Validation of High Precision Web Handling for a Two-Actuator-Based Roll-to-Roll System
Previous Article in Special Issue
Kinect v2-Assisted Semi-Automated Method to Assess Upper Limb Motor Performance in Children
 
 
Article
Peer-Review Record

3D Kinect Camera Scheme with Time-Series Deep-Learning Algorithms for Classification and Prediction of Lung Tumor Motility

Sensors 2022, 22(8), 2918; https://doi.org/10.3390/s22082918
by Utumporn Puangragsa 1, Jiraporn Setakornnukul 1, Pittaya Dankulchai 1 and Pattarapong Phasukkit 2,*
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Sensors 2022, 22(8), 2918; https://doi.org/10.3390/s22082918
Submission received: 22 February 2022 / Revised: 4 April 2022 / Accepted: 9 April 2022 / Published: 11 April 2022
(This article belongs to the Special Issue Kinect Sensor and Its Application)

Round 1

Reviewer 1 Report

General comments:

This study proposes a new motion classification and prediction method using the Kinect v2 device and deep-learning models. Because Kinect V2 is a relatively low-cost technique, the proposed open system may deliver high-quality patient care with reduced costs. In addition, the study applied deep-learning techniques in a clinical problem, and results were evaluated.

The paper is well written except there remain some issues that the author might wish to address before publication. 


Major points:

1. In discussion, please compare deep-learning to other basic regression models (e.g. linear regression, logistic regression, etc.). For instance, the regression model in Ginn et al. seems fine. How do deep-learning models excel the basic regression models?

Ginn, John S., et al. "An image regression motion prediction technique for MRI‐guided radiotherapy evaluated in single‐plane cine imaging." Medical Physics 47.2 (2020): 404-413.

2. Microsoft has discontinued Kinect V2 and the Azure Kinect may be a more up-to-date option. The 3D Kinect technique seems promising for controlling motions in radiotherapy. This study is also useful in other medical imaging applications. Please also discuss the application to other medical devices, e.g. free-breathing cine MRI, PET, etc. 

4. Highlight the limitation of this study. Explain reasons for the lower performance in patients with irregular breathing patterns. Experimental design can be improved by performing a prospective patient study, statistical analysis, etc.  


Minor points:

1. Font size is too small and unclear in figures. Use high-resolution images. 

2. Inconsistent upper- and lower- cases in references. For example, Ref 11, 15, 22 use upper-case letters but others use lower-case letters. Make them consistent. 

3. Inconsistent abbreviated journal names. e.g. some are abbreviated but others are not.  Make them consistent. 

 

Author Response

"Please see the attachment." 

Author Response File: Author Response.pdf

Reviewer 2 Report

The methodology adopted provides a considerable research contribution. However, 100% accuracy in the results still need a proper justification in the results section for each result presented in Table 1, Table 2, Table 3, and Table 4.

General Comments:

The methodology adopted provides a considerable research contribution.

The authors should reconsider the minor rephrasing of the starting sentences of each paragraph.

Specific Comments:

  1. Change the highlighted words for example: "Specifically, of the 400 patient-based input (feature) datasets, 200 datasets (100 each 167 for patients with regular " in Line 167, it can be rephrased as "Specifically, out of the 400 patient-based input ...."
  2. In line 172 : "Meanwhile, there are two groupings of 400 corresponding patient-based output" can be changed with "Meanwhile, there are two groups of 400 corresponding patient-based output ..."
  3. From line 363 to 371, The equations 6 and the corresponding details (line 369-370) should be after line 365 where it the reference of equation 6 is provided, before mentioning the equation 7.
  4. Provide the reference for each mathematical equation where possible for example equation 6.
  5. The fonts in Figure 6, Figure 7, Figure 8, Figure 9 are very small, I suggest increasing the fonts will give better reading ability in printed form.
  6. Page 14, line 497 to 499, Accuracy in 100%, any justification for 100% accuracy ? Similarly, Table 1 in the same page, all results (precision, recall, and F1 score) for classification (0-90%) the results are 100%, please provide justification for it.
  7. Same comment (as mentioned in point 6) for Table 2, where the results have 100% precision , 100% recall, and for F1 Score 100%) and total accuracy is 100% ?
  8. In line 583, again total accuracy is 100% , also in Table 3 ? any justification for 100% accuracy
  9. Same comment for Table 4 (page 19-20).
  10. Conclusion should be accompanied by the results where total accuracies are 100% in line 670.

The authors should reconsider the minor rephrasing of the starting sentences of each paragraph.

Author Response

"Please see the attachment." 

Author Response File: Author Response.pdf

Back to TopTop