Productivity Assessment of the Yolo V5 Model in Detecting Road Surface Damages
Mehmet Ergün Hatır
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe study focuses on the detection of road damage with YOLO, a deep learning method. The method of the study and the findings obtained are noteworthy. However, the introduction section of the manuscript should be improved. Adding the following papers to the study will increase the quality of the manuscript.
· Hatır, E., Korkanç, M., Schachner, A., & Ince, I. (2021). The deep learning method applied to the detection and mapping of stone deterioration in open-air sanctuaries of the Hittite period in Anatolia. Journal of Cultural Heritage, 51, 37-49.
· Hatır, M. E., Ä°nce, Ä°., & Korkanç, M. (2021). Intelligent detection of deterioration in cultural stone heritage. Journal of Building Engineering, 44, 102690.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for Authors1. The motivation and contribution of this paper should be stated more clearly in the abstract to better understand from the beginning of the study. Authors are advised to be precise in the abstract, and structure your abstract as follows- 1) Background 2) Aim/Objective 3) Methodology 4) Results 5) Conclusion. Write 2-4 lines for each and merge everything in one paragraph (200-300 Words) without any subheading.
2. The Introduction section should be improved by adding more recent works in this area and providing a more accurate and informative literature review with the pros and cons of the available approaches and how the proposed method is different comparatively. Also, the motivation and contribution should be stated more clearly.
3. There should be a section for literature review or related work to further give the reason why there is need to the study, the works done in this direction should be review and given their weakness so as to give room for further improvement on the works done. Some latest papers which worked on the similar effects problems should be discussed. i.e review of existing literature on the subject matter should be done.
4. The contribution of the paper should be added at the end of the introduction section for better visibility.
5. The authors need to further emphasize their contributions and relate with the results obtained as is why the adopted methods used are better than the others.
6. Furthermore, the authors should explain in details and explicitly the approach used. The overall description and details regarding how the authors derive the result are not clear.
7. All figures should be explicitly explained and described, only a few variables appearing in them are addressed.
8. The author seems to disregard or neglect some important findings in the results that have been achieved in the paper. So elaborate and explain the results in details.
9. The conclusion is very scanty. The results from the study should be included in the conclusion part.
Comments for author File: Comments.pdf
Although the quality of English used in writing the manuscript is generally good , there are quite a few minor grammatical errors and a careful read through is needed. Spelling mistake should also be corrected.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThis study assesses and contrasts machine learning models utilizing the same dataset of road surface damages gathered from various locations worldwide. By employing two primary evaluation criteria, accuracy and processing speed, this study will showcase the potential of this innovative technology as an exceptionally high-performing model for identifying objects in traffic engineering infrastructure images. However, there are some minor problems in the paper that need further revision:
1. The important experimental data in this paper should be introduced in the abstract.
2. The three headings of the paper are confused and should be revised again.
3. Figures 11 through 15 should be merged into one.
4. Interestingly, the Yolo V5 model, utilizing the SSD detector, yielded an exceptional mAP score of 73.02% and achieved the highest mean at 75.61%, which was quite surprising. please explain why.
5.Explain how to define minor pavement damage and how to assess its impact on model accuracy.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsThe manuscript proposed a technology for the software that can automatically detect road damage, which was considered to produce better results compared to previous models. The study used the same road damage dataset to evaluate and compare machine learning models. By using two main evaluation criteria, accuracy and processing speed, this study demonstrated the potential of the innovative technology as an exceptionally high-performance model for identifying objects in transportation engineering infrastructure. In addition, the findings of this study led to the development of a comprehensive software called RTI-IMS, which can automatically and accurately detect road damage.
The manuscript is not very well organized since the purpose and the significance of the research were not properly demonstrated in sections of Abstract and Introduction. Besides, the numbering of the titles at all levels of the manuscript is incorrect, which significantly affects the readability. In addition, the format of the figures and tables is not standardized which should be carefully revised. Following are some detailed comments that could help the authors improve the manuscript.
Technical Comments:
l Abstract: the abstract should provide a clear and more concise summary of the manuscript, including the adopted research method, main findings, and research significance. The introduction of the research background can be reduced, and the definition of the novel technology employed in the study should be clearly stated. Meanwhile, the “Yolo V5” in the Keywords did not appear in the Abstract which should be revised.
l Introduction: most contents of the Introduction did not touch the main focus of the manuscript, which is the productivity assessment of road surface damage detecting models. The review and comments on current literature are too general rather than indicating the specific research progress and limitations of the damage-detecting models and software. Besides, some references are not strongly related to the cited content, such as the “Gunawan, F. (2018)”.
l Pages 6-9: is it necessary to introduce the “1.1.1. Basic architectures of deep learning” in detail? It is suggested to simplify the content or move part of the content to the Introduction section.
l To demonstrate the superiority of the YOLO V5 model, eight detection models were formed by amalgamating the object detection model with the feature extraction architectures. The method and rationality of forming the eight models are advised to be added to the manuscript.
l The Conclusion of the manuscript seems to be a short summary of the Introduction. It is recommended to make a list of the primary conclusions of the study and provide some important quantitative research results in this section.
Editorial Comments:
l Page 3: “Two (02) most important evaluation criteria of a machine learning detection model are accuracy and speed…” why do the authors use “(02)” here?
l Page 11: Figure 7 has a bad presentation. The definitions of different shapes and connecting lines are suggested to be added, and the clarity of the image should be improved.
Comments for author File: Comments.pdf
Extensive editing of English language required
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 5 Report
Comments and Suggestions for Authors· In the abstract, the authors are recommended to display the basic methodology and findings in a quantitative manner.
· Research innovation should be stated at the end of previous studies. The research gap is not clear.
· The research objectives are not clear.
· The literature reviewing should be summarized to be concise. Please write it as a whole paragraph not separate ones.
· There many other technique can be applied in determining pavement damage like laser scanning. The findings of this study can be used in comparison (Applications of Terrestrial Laser Scanner in Detecting Pavement Surface Defects, https://doi.org/10.3390/pr11051370)
· The authors should rethink about the display method. The manuscript looks like a project report.
· The manuscript should be shorten and the discussion should be enhanced by comparing the findings with the previous studies.
· The authors should give their findings in a quantitative manner.
· Many figures can be deleted (figure 9, 7, 4).
· The findings of the proposed model should be compared with other studies.
· Recommendations to the practitioners should be added. In addition, a limitation of the current study should be highlighted.
· In the abstract, the authors are recommended to display the basic methodology and findings in a quantitative manner.
· Research innovation should be stated at the end of previous studies. The research gap is not clear.
· The research objectives are not clear.
· The literature reviewing should be summarized to be concise. Please write it as a whole paragraph not separate ones.
· There many other technique can be applied in determining pavement damage like laser scanning. The findings of this study can be used in comparison (Applications of Terrestrial Laser Scanner in Detecting Pavement Surface Defects, https://doi.org/10.3390/pr11051370)
· The authors should rethink about the display method. The manuscript looks like a project report.
· The manuscript should be shorten and the discussion should be enhanced by comparing the findings with the previous studies.
· The authors should give their findings in a quantitative manner.
· Many figures can be deleted (figure 9, 7, 4).
· The findings of the proposed model should be compared with other studies. Discuss the accury of the proposed model
· Recommendations to the practitioners should be added. In addition, a limitation of the current study should be highlighted.
Comments on the Quality of English LanguageShould be improved
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 4 Report
Comments and Suggestions for Authors1. Abstract: Compared to the previous manuscript, the research methods, main findings, and research significance are better stated in the abstract. However, the appearance of evaluation indicators is too abrupt, and a brief explanation can be given as to why accuracy (mAP) and processing speed were chosen as evaluation indicators.
2. Introduction: In response to the reviewer's comments, comments on the research progress and limitations of the loss detection models and software in the current literature were added in the introduction.
3. The description of the formation methods of the 8 detection models is relatively complete in the revised manuscript, but the demonstration of rationality is not prominent enough.
4. The conclusion still lacks some important quantitative results that the RIT-IMS software can be obtained.
5. Please remove the circles in front of Equations (1) ~ (2), (4) ~ (6).
6. Why did the authors leave a blank line before the second paragraph in section “5.3.2. Model processing speed”?
7. Tables 2~5 have inconsistent fonts, and Table 5 is incomplete.
8. The layout, format, and spacing of Figure 12 and Figure 5 (on page 25) are chaotic, and Figure 5 on page 25 should be Figure 13. Besides, the text in the two figures overlaps. The authors are strongly recommended to check and revise the format of the whole manuscript.
Comments on the Quality of English LanguageModerate editing of English language required.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 5 Report
Comments and Suggestions for AuthorsThe authors addressed the reviewer's comments
Comments on the Quality of English Languagecan be improved
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 3
Reviewer 4 Report
Comments and Suggestions for AuthorsNo further comments for the authors.
Comments on the Quality of English LanguageMinor editing of English language required.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
