Deep Learning-Based Human Body Posture Recognition and Tracking for Unmanned Aerial Vehicles
Round 1
Reviewer 1 Report
The article is very interesting. It has a high scientific contribution. The methodological procedures are robust. The writing is clear and easy to understand. Recommend only as references with some article from 2022.Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
This research article describes a method for estimating an automated system of people pose recognition by deep learning technics. To do so, authors uses UAV system for image adquisition. The design and development of a new remotely piloted aircraft system is proposed. But nowhere in the text does the justification for this new development appear. It has no relation with the title of the article nor with the subject of the research. I would ask the authors to justify why this new UAV design is necessary or remove this part of this research article.
Table 4 caption does not describe table content
Figure 20 is out of focus. Not understandable.
Figure 25 caption seems incorrect.
After reading this research article, it is not clear to me where the image processing and identification takes place. Is it done after the transmission of the images on another computer, server, etc. or in real time? I kindly ask the authors to clarify this.
Lines 342-343: It is not clear to the reader what the authors refer to by the Hungarian algorithm. Please describe it, with references if necessary.
Lines 395, 429, 457: Although MobileNet, Alexnet and VGGNet are correctly described, there are no citations or references to them.
Lines 546-551: Nowhere in the text does it appear how the reference trajectory has been established for calculation and comparison.
Lines 575-578: I would suggest to enumerate the possible body poses as bullet point list, in order to make it easier to readers.
Lines 646-658: Fuzzy logic are not well described. It does not appear where this analysis is done or how have been implemented. I would suggest to authors to rewrite this paragraph because is unclear and hard to read.
Lines 579-580: the same as 575-578
Lines 625: I would be grateful if the authors could briefly describe what each metric means and what their optimal values are.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
Dear Authors,
Thank you for your work editing and answering all my questions and doubts.
Nevertheless, I have to confirm my first reviewing point: the development of a new type or improvement of UAV design is not justified in the text. You point out that the stable maintenance of the UAV yaw must be kept in equilibrium during flight operation. This is not justified as necessary in the title, abstract or text. Any UAV operation equipped with gimball can keep the object in focus without additional mechanical equipment.
You point out that yaw must be maintained, but you do not mention the need to maintain pitch or roll. These movements would also distort the acquisition of images of the target object. Why the yaw and not the other displacements?
I kindly invite the authors to justify the section on new UAV design, or restructure the title, abstract orthe whole article, as I consider that, in the current state of revision, this UAV section is not justified.
Regarding the other comments, I can not follow your changes and indications because the lines indicated do not correspond with those in the pdf file. Please indicate me the changes with the correct line numbers.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
