Applicability of Machine Learning for Vessel Dimension Survey with a Minimum Number of Common Points

Round 1

Reviewer 1 Report

The paper is devoted to the problems that arise when controlling the size of ships, platforms, and offshore installations. The authors proposed an approach for transforming three-dimensional spatial coordinates based on machine learning methods when only 3 common points are known. The proposed methodology is to find and apply a simple method of compacting the number of points in less geodetic networks, checking the possibility of using an artificial neural network to solve the problem of transforming coordinates for marine purposes using a point cloud thus created, checking whether the applied approach allows achieving the same or higher 3D transformation accuracy compared to traditionally used computational methods. The results show that the proposed method provides equal or better accuracy than the commonly used conformal transformations.

 

Despite the satisfactory quality of the article, there are some shortcomings that need to be corrected.

1. The abstract should be expanded with obtained results.
2. The object of the study should be described.
3. The data used for the experimental study should be described in detail.
4. The purpose of selection described architecture of artificial neural network should be described.
5. The division of the sample in proportions the training sample constituting 70% of the total data, the validation sample 15%, and the test sample 15% should be justified.
6. It should be described why a dataset of 105 observations is sufficient to obtain a reasonable result.
7. It is recommended to divide figure 6 into three figures and increase their size for better interpretation.
8. It is recommended to divide figure 7 into four figures and increase their size for better interpretation.
9. The Discussion section should be included and numerical results should be compared with other methods and approaches.

In summarizing my comments I recommend that the manuscript is accepted after major revision.

Author Response

Answers for comments of the review 1 of manuscript ID: applsci-1603682

 

Title:  Applicability of machine learning for vessel dimension survey with a minimum number of common points

 

Dear Reviewer,

Thank you very much for a great deal of work in reviewing our manuscript. Thank you for your comments and suggestions for improving the manuscript. We have tried to complete all of them according to your suggestions. The following amendments and changes were made:

1. The abstract should be expanded with obtained results.

The abstract was expanded and complemented by the results achieved.

2. The object of the study should be described.

The research objective is clarified and described in a broader context in the introduction section as suggested.

3. The data used for the experimental study should be described in detail.

The description of the data used in the study has been included in the Materials and Methods section.

4. The purpose of selection described architecture of artificial neural network should be described.

The Materials and Methods section has been expanded by a description of how the neural network architecture has been selected.

5. The division of the sample in proportions the training sample constituting 70% of the total data, the validation sample 15%, and the test sample 15% should be justified.

The division of data to training and validation group  was justified and the reference to the literature was added.

6. It should be described why a dataset of 105 observations is sufficient to obtain a reasonable result.

In the justification, formula (3) was applied, based on which the minimum number of points for training neural networks was calculated. The value of 105 slightly exceeds the required number of points.

7. It is recommended to divide figure 6 into three figures and increase their size for better interpretation.

The image size and resolution has been increased. Each picture is now presented separately.

8. It is recommended to divide figure 7 into four figures and increase their size for better interpretation.

The image size and resolution has been increased. Each picture is now presented separately.

 

Moreover:

1. The comments and suggestions of other reviewers have been taken into account.
2. The file with all the changes as follow-up changes, has been attached in pdf format.
3. The English editing certificate has been attached in pdf format.

 

Once again, we would like to thank you very much for your deep insight into the review of our manuscript. It has greatly helped us to improve our manuscript.

 

                                                                     Authors

Author Response File: Author Response.docx

Reviewer 2 Report

Good article, well focused, but limited in its approach. In the opinion of the reviewer the study would gain a lot with more applications for the industry (off shore industry is cited) but more projection would be welcome. A field of great potential for applied engineering.

Author Response

Answer on the review 2 of manuscript ID: applsci-1603682

 

Title:  Applicability of machine learning for vessel dimension survey with a minimum number of common points

 

Dear Reviewer,

Thank you very much for a great deal of work you did in reviewing our manuscript. Thank you for your comments and suggestions for improving the manuscript. We have tried to complete all of them according to your suggestions. The following amendments and changes have been made:

1. More projection would be welcome.

Yes, it would be interesting to use more projections. That is what we are planning in the future.  We addressed it in the text.

2. A field of great potential for applied engineering.

More attention has been paid to the potential application of the method in other areas of civil engineering (in the Introduction and Conclusions sections).

 

Moreover:

1. The comments and suggestions of other reviewers have been taken into account.
2. The file with all the changes as follow up changes, has been attached in pdf format.
3. The English editing certificate has been attached in pdf format.

 

Once again, we would like to thank you very much for your great input into the review of our manuscript. It has greatly helped us to improve our manuscript.

 

Authors

Author Response File: Author Response.docx

Reviewer 3 Report

Overall speaking, this is a high quality manuscript. It is well written. 

I only have one doubt. In page 2, line 80 and line 384, it is claimed that this method presented in this manuscript has better or equal accuracy than traditional method. I am wondering

1) Did you make the comparison to other NOVEL method published in recent years instead of the TRADITIONAL ones?

2) The claim was verified to a certain extend in this manuscript. However, if this same method is applied to a broad scenario, does it stand out better accuracy as well? Why and how? Could you explain?

 

Author Response

Answer on the review 3 of manuscript ID: applsci-1603682

 

Title:  Applicability of machine learning for vessel dimension survey with a minimum number of common points

 

Dear Reviewer,

Thank you very much for the great deal of work you did in reviewing our manuscript. Thank you for your comments and suggestions for improving the manuscript. We have tried to complete all of them according to your suggestions. Especially the following amendments and changes were made:

1. In page 2, line 80 and line 384, it is claimed that this method presented in this manuscript has better or equal accuracy than the traditional method. I am wondering. Did you make the comparison to other NOVEL method published in recent years instead of the TRADITIONAL ones.

The Introduction and Discussion sections refer to the results obtained with gradient methods, which give very similar results to the similarity transformation with the least-squares method.

2. The claim was verified to a certain extend in this manuscript. However, if this same method is applied to a broad scenario, does it stand out better accuracy as well? Why and how? Could you explain?

In the Discussion and Conclusions sections, a limited application of the method was indicated. This limitation applies to extrapolation scenarios where points are recalculated outside the neural network learning domain.

Moreover:

1. The comments and suggestions of other reviewers have been taken into account.
2. The file with all the changes as follow up changes, has been attached in pdf format.
3. The English editing certificate has been attached in pdf format.

 

Once again, we would like to thank you very much for your great input into the review of our manuscript. It has greatly helped us to improve our manuscript.

Authors

Author Response File: Author Response.docx

Reviewer 4 Report

The article addresses interesting challenges in dimensional control of floating objects and a novel approach that uses machine learning to measure ship dimensions with 3D spatial coordinate transformations. The methodology used is sound and the results are interesting and provide useful insights.

The paper is well structured and clearly written and can be easily followed by readers who are not necessarily familiar with the measurement of ship dimensions. The topic is current and interesting.

However, there are also some minor weaknesses that detract from the final appearance of the paper:

1. applied sciences/MDPI Instructions for authors should be followed closely:
2. Acronyms/abbreviations/initialisms should be defined when they first appear in each of the three sections: in the abstract; in the main text; in the first figure or table. When defined for the first time, the acronym/abbreviation/initialism should be added in brackets after the spelled out form (e.g. line 13, 20, etc.)
3. Line 27, 29, 32 etc. - References should be cited, such as. [2-4].
4. The quality of figures 5-9 is rather low.

Author Response

Answer on the review 4 of manuscript ID: applsci-1603682

 

Title:  Applicability of machine learning for vessel dimension survey with a minimum number of common points

 

Dear Reviewer,

Thank you very much for the great deal of work you did in reviewing our manuscript. Thank you for your comments and suggestions for improving the manuscript. We have tried to complete all of them according to your suggestions. The following amendments and changes were made:

1. Applied Sciences/MDPI Instructions for authors should be followed closely:

Acronyms/abbreviations/initialisms should be defined when they first appear in each of the three sections: in the abstract; in the main text; in the first figure or table. When defined for the first time, the acronym/abbreviation/initialism should be added in brackets after the spelled out form (e.g. line 13, 20, etc.)

The use of abbreviations and acronyms has been improved in accordance with the instructions to authors and suggestions.

Line 27, 29, 32 etc. - References should be cited, such as. [2-4].

It has been improved in accordance with the instructions to authors and suggestions.

1. The quality of figures 5-9 is rather low.

The image size and resolution have been increased. Each picture is now presented separately.

 

Moreover:

1. The comments and suggestions of other reviewers have been taken into account.
2. The file with all the changes as follow up changes, has been attached in pdf format.
3. The English editing certificate has been attached in pdf format.

 

Once again, we would like to thank you very much for your significant input into the review of our manuscript. It has greatly helped us to improve our manuscript.

Authors

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Thanks for the authors for considering the reviewer's comments and recommendations. In my opinion, now the article can be accepted.