Cross-Perspective Human Behavior Recognition Based on a Joint Sparse Representation and Distributed Adaptation Algorithm Combined with Wireless Optical Transmission
Round 1
Reviewer 1 Report
Dear Authors,
It would be appropriate to make the following major corrections.
1. The summary is too long. Contains many background details. It should describe the project, methodology and results.
2. The use of Optical Zoom Antenna variants for wireless communication is not a new method.
3. The aim of the article is to recognize human behavior and the purpose of wireless optical communication cannot be understood. Visible field communication is a standard optical communication type and is not a novelty. It may therefore be referenced instead of section 2.2.
4. The contribution to the algorithm is not fully understood. The place and comparison of the algorithm in the literature should be made.
Author Response
It would be appropriate to make the following major corrections.
- The summary is too long. Contains many background details. It should describe the project, methodology and results.
Answer: I have revised the abstract as required.
- The use of Optical Zoom Antenna variants for wireless communication is not a new method.
Answer: I have deleted this part of the instructions as required.
- The aim of the article is to recognize human behavior and the purpose of wireless optical communication cannot be understood. Visible field communication is a standard optical communication type and is not a novelty. It may therefore be referenced instead of section 2.2.
Answer: I have revised this part as required.
- The contribution to the algorithm is not fully understood. The place and comparison of the algorithm in the literature should be made.
Answer: In Section 3.2, I compare and illustrate the algorithm.
Reviewer 2 Report
This paper attempts to tackle cross perspective human behaviour recognition (abbreviate as CP-HBR), which is indeed an interesting and real-world applicable task. However, the paper is poorly presented, and few to none useful information is provided for the development of the discussed topic. I list several concerns as follows:
1. The paper did review the joint sparse representation and distribution adaptation (more commonly known as domain adaptation), but did not state the reason such representation is leveraged for CP-HBR task. A method that is useful towards the research/industry community should state why such method is proposed at the first place, so that readers could understand the purpose of the design. Otherwise, it seems that the methods are leveraged ONLY because they present decent results. It should also be pointed out that neither joint sparse representation nor distribution adaptation is a novel method for behaviour recognition. If the authors intend to stress that they are novel for the cross perspective task, they should state it clearly with solid review of prior works.
2. The paper states from time to time the use of wireless communication optical technology. Unfortunately, after reading the whole paper I cannot understand the role wireless communication play in the method. If it is leveraged merely for data collection, it should be stated very clearly and not mixed with the data processing module (i.e., the method introduced in Section 2.3). If not, it should be much clearly stated how the wireless communication technology plays a role in the designed method. The authors should understand that the purpose of a paper is for readers/researchers to understand the system/method proposed, not to just showcase a set of techniques which may not seemingly be correlated.
3. The presentation of the paper is poor. While there are a lot of formula, it is almost impossible to follow them easily since only a handful of explanation is provided for the formula. This essentially deteriorate the quality of the presentation. Moreover, the figures are also of low quality and not informative. The fonts in the figures are either too big or too small, and the captions are really as simple as it could be (almost to the extend that it could be negligible). The presentation is unprofessional.
Given the unclear motivation, lack of novelty, unprofessional and ambiguous presentation, this paper should be rejected.
Author Response
This paper attempts to tackle cross perspective human behaviour recognition (abbreviate as CP-HBR), which is indeed an interesting and real-world applicable task. However, the paper is poorly presented, and few to none useful information is provided for the development of the discussed topic. I list several concerns as follows:
- The paper did review the joint sparse representation and distribution adaptation (more commonly known as domain adaptation), but did not state the reason such representation is leveraged for CP-HBR task. A method that is useful towards the research/industry community should state why such method is proposed at the first place, so that readers could understand the purpose of the design. Otherwise, it seems that the methods are leveraged ONLY because they present decent results. It should also be pointed out that neither joint sparse representation nor distribution adaptation is a novel method for behaviour recognition. If the authors intend to stress that they are novel for the cross perspective task, they should state it clearly with solid review of prior works.
Answer: I have explained in the abstract the role and advantages of joint sparse representation and distributed adaptive techniques in the field of human behavior recognition across perspectives.
- The paper states from time to time the use of wireless communication optical technology. Unfortunately, after reading the whole paper I cannot understand the role wireless communication play in the method. If it is leveraged merely for data collection, it should be stated very clearly and not mixed with the data processing module (i.e., the method introduced in Section 2.3). If not, it should be much clearly stated how the wireless communication technology plays a role in the designed method. The authors should understand that the purpose of a paper is for readers/researchers to understand the system/method proposed, not to just showcase a set of techniques which may not seemingly be correlated.
Answer: I have explained the wireless communication optical technology in the introduction as required.
- The presentation of the paper is poor. While there are a lot of formula, it is almost impossible to follow them easily since only a handful of explanation is provided for the formula. This essentially deteriorate the quality of the presentation. Moreover, the figures are also of low quality and not informative. The fonts in the figures are either too big or too small, and the captions are really as simple as it could be (almost to the extend that it could be negligible). The presentation is unprofessional.
Answer: I have modified the formula and the diagram as suggested.
Reviewer 3 Report
The paper proposes a scheme for the recognition of human behavior (based on an algorithm that combines joint sparse representation and distributed adaptation) as well as the transmission of data through an optical wireless link.
General comments
I think that the main drawback of the paper is that associates two different aspects, the recognition algorithm and the mode of data transmission (a wireless optical link). The algorithm does include novel material (and as such is worth publishing) however the choice of wireless optical technology does not really add to the paper’s overall value. For example: What does this particular technology offer compared to the fiber-optic one? Given the distances shown in Table 1, a fiber link (even with a multimode fiber) could provide a secured and reliable mode of transmission.
Within the above framework, I would propose the following changes:
Ø In the title, the phrase “in Wireless Communication Optical Technology” should be deleted
Ø The abstract should be re-written and include the part from line 24 or 25 onwards. If the authors think that they should refer to wireless optical technology, they should do it through only a couple of lines at the end of the abstract.
Ø Lines 85-98, and subsections 2.2 and 3.1 should be either removed or become a small subsection making a short reference to wireless optical technology as a possible mode of data transmission.
I am not sure if the references in the “1. Introduction” section are properly cited.
The paper needs extensive editing regarding the use of English (mainly the way of expression).
Specific comments
Clarifications are needed regarding eqs. (2) to (5). For example, is A a mxn matrix? a1, a2, … an are mx1 column vectors? How is W Î Rmxn when he has k elements? Is Wi a row or a column vector? Why is the last element index denoted with nj and not with n?
The phrases “much higher” in line 31 (abstract) and “significantly lower” in line 357 are a bit exaggerated. Improvement in recognition accuracy does exist (and it is welcome) but in most cases is between 2% and 6%.
In “4. Conclusions” (last sentence) the authors mention “problems and deficiencies in this paper”. Evidently, the authors refer to issues for further research, so the words “problems” and “deficiencies” are rather negative (and unfair) for their work. The authors should simply make a short reference to such issues and consider them as topics for further investigation.
Reviewing decision
Overall, the paper is publishable (mainly due to the material regarding the combination of the joint sparse representation and the distributed adaptation algorithms) subject to the comments made above.
Author Response
The paper proposes a scheme for the recognition of human behavior (based on an algorithm that combines joint sparse representation and distributed adaptation) as well as the transmission of data through an optical wireless link.
General comments
I think that the main drawback of the paper is that associates two different aspects, the recognition algorithm and the mode of data transmission (a wireless optical link). The algorithm does include novel material (and as such is worth publishing) however the choice of wireless optical technology does not really add to the paper’s overall value. For example: What does this particular technology offer compared to the fiber-optic one? Given the distances shown in Table 1, a fiber link (even with a multimode fiber) could provide a secured and reliable mode of transmission.
Answer: I have explained the advantages of wireless optical technology.
Within the above framework, I would propose the following changes:
Ø In the title, the phrase “in Wireless Communication Optical Technology” should be deleted
Ø The abstract should be re-written and include the part from line 24 or 25 onwards. If the authors think that they should refer to wireless optical technology, they should do it through only a couple of lines at the end of the abstract.
Answer: I have added the description of wireless optical technology at the end of the summary.
Ø Lines 85-98, and subsections 2.2 and 3.1 should be either removed or become a small subsection making a short reference to wireless optical technology as a possible mode of data transmission.
Answer: I have deleted section 2.2.
I am not sure if the references in the “1. Introduction” section are properly cited.
Answer: The quote section is correct.
The paper needs extensive editing regarding the use of English (mainly the way of expression).
Answer: I have modified the English use in the text.
Specific comments
Clarifications are needed regarding eqs. (2) to (5). For example, is A a mxn matrix? a1, a2, … an are mx1 column vectors? How is W Î Rmxn when he has k elements? Is Wi a row or a column vector? Why is the last element index denoted with nj and not with n?
Answer: I have given a simple explanation of the formula.
The phrases “much higher” in line 31 (abstract) and “significantly lower” in line 357 are a bit exaggerated. Improvement in recognition accuracy does exist (and it is welcome) but in most cases is between 2% and 6%.
Answer: I have revised the description as required.
In “4. Conclusions” (last sentence) the authors mention “problems and deficiencies in this paper”. Evidently, the authors refer to issues for further research, so the words “problems” and “deficiencies” are rather negative (and unfair) for their work. The authors should simply make a short reference to such issues and consider them as topics for further investigation.
Answer: I have explained these problems and deficiencies in the conclusion.
Reviewing decision
Overall, the paper is publishable (mainly due to the material regarding the combination of the joint sparse representation and the distributed adaptation algorithms) subject to the comments made above.
Round 2
Reviewer 1 Report
Dear Author,
The suggested corrections have been made. However, the following work can also be referenced regarding wireless optical communication:
Murat Yucel and Muharrem Acikgoz, "Optical Communication Infrastructure in New Generation Mobile Networks", FIBER AND INTEGRATED OPTICS, 2023, doi.org/10.1080/014680300.2023.2186811.
Regards.
Author Response
Comments and Suggestions for Authors
Dear Author,
The suggested corrections have been made. However, the following work can also be referenced regarding wireless optical communication:
Murat Yucel and Muharrem Acikgoz, "Optical Communication Infrastructure in New Generation Mobile Networks", FIBER AND INTEGRATED OPTICS, 2023, doi.org/10.1080/014680300.2023.2186811.
Answer: I have already quoted the literature in the article.
Reviewer 2 Report
I thank the reviewers for their attempt in clarifying the motivation of the paper and its relation with the wireless communication technology. Although the techniques proposed for cross-perspective human behavior recognition are NOT novel, its application is worth research and this paper contributes to the research community. Technically, the paper is now more acceptable than the previous form.
However, it should be pointed out that there are various formatiing and typo issues that hinders the readability of the work. The authors are suggested to go through a thorough editing of the paper for the format to be more professional.
Author Response
Comments and Suggestions for Authors
I thank the reviewers for their attempt in clarifying the motivation of the paper and its relation with the wireless communication technology. Although the techniques proposed for cross-perspective human behavior recognition are NOT novel, its application is worth research and this paper contributes to the research community. Technically, the paper is now more acceptable than the previous form.
However, it should be pointed out that there are various formatiing and typo issues that hinders the readability of the work. The authors are suggested to go through a thorough editing of the paper for the format to be more professional.
Answer: I've already revised the article.
Reviewer 3 Report
Please, see submitted file.
Comments for author File: 
Comments.pdf
Author Response
Checks by reviewers: I still think it's best to stick to the behavior recognition algorithm for manuscripts, and don't associate it with a specific mode of communication. If the author insists on citing optical wireless technology, I will accept it, but I prefer the title that emphasizes algorithm rather than transmission technology. In this regard, I would like to propose a title, such as "Cross-perspective human behavior recognition based on joint sparse representation and distributed adaptive algorithm combined with wireless optical transmission".
Answer: I have revised the title of the article as required.
Within the above framework, I suggest the following changes:
The phrase "in optical technology of wireless communication" in the title should be deleted.
The abstract should be rewritten and include the part starting from line 24 or 25. If the authors think they should cite wireless optical technology, they should write a few lines at the end of the abstract.
Answer: I added a description of wireless optical technology at the end of the summary.
Examiner's inspection:
For the title, please see the comments above.
Answer: I have revised the title of the article as required.
As for the abstract, it should be shorter and only include information related to the work described in the paper (not background information). I think the summary based on the 30th to 40th behaviors (revised edition) is enough.
Answer: I have revised the abstract and shortened the number of words in the abstract.
Lines 85-98 and subsections 2.2 and 3.1 should be deleted or become a small section, which briefly mentions wireless optical technology as a possible data transmission mode.
A: I have deleted section 2.2.
Examiner's check: OK.
I'm not sure "1. The "Introduction" section is quoted appropriately.
Answer:The quotation is correct.
Examiner's check: The problem still exists. According to the policy of periodicals, references are usually cited by number (such as [1]) or author's surname (such as Chien). In manuscripts, references are cited in two ways (numbers and surnames).
Answer: I'm sorry. Now, I have understood the modification requirements and modified them.
Regarding the use of English (mainly expression), this paper needs a lot of editing.
Answer: I revised the English usage in the text.
Check of the reviewers: I am not a native speaker myself, but I think it needs further editing (mainly about expression).
Answer: I've already revised the article.
Specific opinions
Eqs needs to be clarified. (2) to (5). For example, is A an mxn matrix? Are a1, a2, … an mx1 column vectors? What about when wrmnxn has k element? Is Wi a row vector or a column vector? Why is the last element index represented by nj instead of N?
Answer:I have made a simple explanation of the formula.
Examiner's check: I can still see the problems in equations (2) to (5). For example, if a is a line vector (1xn because it is seen) and xirmxn, then it is impossible to multiply Ax. The same applies to Wx. Besides, if wrmnx. Then why does it have k elements, and why is the last element index represented by nj instead of N?
Answer: I have understood the crux of the problem and revised the formula and its explanation.
The phrase "much higher" in line 31 (abstract) and "significantly lower" in line 357 are somewhat exaggerated. The improvement of recognition accuracy does exist (and is welcome), but in most cases it is between 2% and 6%.
Answer: I have revised the instructions as required.
Examiner's check: OK.
In "4. Conclusion (the last sentence) The author mentioned "the problems and shortcomings of this article". Obviously, the author mentioned the problems that need further study, so "problems" and "defects" are quite negative (and unfair) for their work. The author should simply mention these problems and regard them as the subject of further study.
Answer: I have explained these problems and shortcomings in my conclusion.
Reviewer's check: I don't agree to use the word "defect". A research work may leave unsolved problems (which may be the subject of further work), but if a paper has "problems and shortcomings", it means that it is not well written. I suggest that the author use such a phrase.
As "unsolved problems"/"problems to be further studied", not words such as "problems" or "deficiencies".
Answer: I have revised it and used more appropriate instructions.
Review decision
The manuscript needs further revision (according to the comments written in red above) before it can be published.
Round 3
Reviewer 3 Report
I still see issues in eqs (1) to (5) that need clarification.
For example:
1. I gather that a1, a2, …an in A are row vectors themselves. If they really are, the authors should say so and A becomes a matrix rather than a column vector.
2. Why does the last element in eq. (2) have a subscript nj? Shouldn’t that be “n” (since W is an Rmxn matix)?
3. Why does the last element in eq. (3) have also a subscript nj? Shouldn’t that be “m”? If W is an Rmxn matix, then Wi should have m rows.
4. In eq (4), “y” appears as a single element. However, according to (5), “y” should be a mx1 column vector.
I would really appreciate it if the authors could answer to the above issues one by one in order to become clear whether it is only a matter of clarifications or modifications in the eqs are in order.
I am satisfied with the rest of the authors’ response.
I will still state that the paper needs major revision only in case modifications in eqs. (1) to (5) are needed.
Author Response
Comments and Suggestions for Authors
I still see issues in eqs (1) to (5) that need clarification.
For example:
- I gather that a1, a2, …an in A are row vectors themselves. If they really are, the authors should say so and A becomes a matrix rather than a column vector.
Answer: A is indeed A matrix, and its elements are row vectors, which I have modified.
- Why does the last element in eq. (2) have a subscript nj? Shouldn’t that be “n” (since W is an Rmxn matix)?
Answer: I have modified it, and the element subscript is indeed n.
- Why does the last element in eq. (3) have also a subscript nj? Shouldn’t that be “m”? If W is an Rmxn matix, then Wi should have m rows.
Answer: I have modified it, and the element subscript is indeed m.
- In eq (4), “y” appears as a single element. However, according to (5), “y” should be a mx1 column vector.
Answer:Yes, y is indeed a column vector. I am very grateful for your comments, which is of important guiding value for the work of this article.
I would really appreciate it if the authors could answer to the above issues one by one in order to become clear whether it is only a matter of clarifications or modifications in the eqs are in order.
I am satisfied with the rest of the authors’ response.
I will still state that the paper needs major revision only in case modifications in eqs. (1) to (5) are needed.
Round 4
Reviewer 3 Report
I am satisfied with the authors’ response. The paper is publishable subject to minor revision (minor to moderate scale editing regarding the use of English, e.g. lines 164-166).
Author Response
Comments and Suggestions for Authors
I am satisfied with the authors’ response. The paper is publishable subject to minor revision (minor to moderate scale editing regarding the use of English, e.g. lines 164-166).
Reply: Thanks for your kindly comments. I have carefully made a editing regarding the use of English from beginning to end.
