Next Article in Journal
A Blockchain Storage Architecture Based on Information-Centric Networking
Previous Article in Journal
Smart, Fast, and Low Memory Beam-Steering Antenna Configurations for 5G and Future Wireless Systems
 
 
Article
Peer-Review Record

Improved YOLOv5 Based on Hybrid Domain Attention for Small Object Detection in Optical Remote Sensing Images

Electronics 2022, 11(17), 2657; https://doi.org/10.3390/electronics11172657
by Tianmin Deng *, Xuhui Liu and Guotao Mao *
Reviewer 1:
Reviewer 2:
Reviewer 3:
Electronics 2022, 11(17), 2657; https://doi.org/10.3390/electronics11172657
Submission received: 9 July 2022 / Revised: 20 August 2022 / Accepted: 21 August 2022 / Published: 25 August 2022

Round 1

Reviewer 1 Report

Point 1 : Change the font in Figure 1 to normal to increase visibility.

Point 2 :    Increase the resolution so that Figure 2 can be seen clearly.

Point 3 :   Make the fonts in Figures 3 and 4 the same as in Figure 5 to increase visibility.

Point 4 :  Add average detection time to Table 1.

Point 5 :  Delete lines 412 (a) through (f).

Author Response

Point 1: Change the font in Figure 1 to normal to increase visibility.

 Response 1: We are grateful for the suggestion. In order to see the content in Figure 1 more clearly, we have changed the font in Figure 1 to normal.

Point 2: Increase the resolution so that Figure 2 can be seen clearly.

Response 2: Thank you for your advice. According to the reviewer's suggestion, we have increased the resolution of Figure 2.

Point 3: Make the fonts in Figures 3 and 4 the same as in Figure 5 to increase visibility.

Response 3: We are very grateful for the suggestion. According to reviewer's comment, we have made the fonts in figures 3 and 4 the same as those in Figure 5 to improve visibility.

Point 4: Add average detection time to Table 1.

Response 4: We apologize for this problem in the manuscript. We have added the average detection time of each image to the evaluation metrics. And a column of average detection time has been added in Table 1.

Point 5: Delete lines 412 (a) through (f).

Response 5: We apologize for this problem in the manuscript. We have deleted lines 412-428 because there are some errors in the typesetting.

Reviewer 2 Report

1- Abstract need to be improved to cover the objectives of paper.

2- Introduction section is very long, try to be more concise. Divide your section into four paragraphs: The importance of water security, the previous studies and the methods used, the problem statement (research gaps), and the novelty of this research and the objectives.

3-The discussion section needs to be more specific in comparison with the previous researchers. Please improve it by adding new references and polish your results and the differences the results obtained and the previous findings.

4- English language needs more improvements to be qualified for publication.

Author Response

Response to Reviewer 2 Comments

Point 1: Abstract need to be improved to cover the objectives of paper.

 Response 1: Thank you very much for your suggestion. The objective of this paper is to solve the problem of small target detection in optical remote sensing images. We have improved the abstract and covered the objectives of the paper.

Point 2: Introduction section is very long, try to be more concise. Divide your section into four paragraphs: The importance of water security, the previous studies and the methods used, the problem statement (research gaps), and the novelty of this research and the objectives.

Response 2: We appreciate your suggestion. In order to be clearer and meet the concerns of the reviewer, we have divided the introduction into four paragraphs: the importance of remote sensing target detection, previous research and methods used, problem statements, and the novelty and objectives of remote sensing small target detection. And expand in the revised manuscript to introduce in detail.

Point 3: The discussion section needs to be more specific in comparison with the previous researchers. Please improve it by adding new references and polish your results and the differences the results obtained and the previous findings.

 Response 3: We are very grateful to the reviewer for the suggestion. According to the reviewer's comment, in the discussion section, we have added new references [30] and [31]. The data of YOLOv4, DFPN-YOLO and T-TRD-DA have been supplemented in Table Ⅱ. We have polished our results and the differences between the results and the previous findings in the revised manuscript.

Point 4: English language needs more improvements to be qualified for publication.

Response 4: We apologize for the language problems in the original manuscript. We have further improved the English language.

Reviewer 3 Report

The tile is quite interesting, but have some concerns.

1. The introduction section is missing some details like motivation, novelty, and contributions. It is advisable to add separate subsection for these.

2. In related works section, the authors must add a relative comparison of the proposed approach with the state-of-the-art approaches for better clarity to the readers.

3. Quality of all figures are poor. Redraw as per the journal guidelines.

4. Equations must be cited with a relevant source (if referred from some paper)

5. Line number 411, something is missing. Complete it.

6. Conclusion must specify the findings of the paper in numerical terms and clearly specify the future work.

Author Response

Response to Reviewer 3 Comments

Point 1: The introduction section is missing some details like motivation, novelty, and contributions. It is advisable to add separate subsection for these.

Response 1: We deeply appreciate the reviewer’s suggestion. In order to be clearer and meet the concerns of the reviewer, we have divided the introduction into four paragraphs: the importance of remote sensing target detection, previous research and methods used, problem statements, and the novelty and objectives of remote sensing small target detection. Among them, the first paragraph is motivation, and the fourth paragraph includes novelty and contribution. We have introduced it in detail in the revised manuscript.

Point 2: In related works section, the authors must add a relative comparison of the proposed approach with the state-of-the-art approaches for better clarity to the readers.

Response 2: We are grateful for the suggestion. To be more clear and in accordance with the reviewer concerns, first of all, we have added some of the state-of-the-art approaches in the revised manuscript. Secondly, we have compared the methods in this paper with the most advanced methods. Please refer to related works for details.

Point 3: Quality of all figures are poor. Redraw as per the journal guidelines.

Response 3: We are extremely grateful to reviewer for pointing out this problem. We have redrawn all the pictures in the revised manuscript according to the journal guidelines.

Point 4: Equations must be cited with a relevant source (if referred from some paper)

Response 4: Thank you very much for your suggestion. The equations of hybrid domain attention unit for feature extraction refer to references [25] and [26], which have been marked in the paper.

Point 5: Line number 411, something is missing. Complete it.

Response 5: We apologize for this problem in the manuscript. We have deleted lines 412-428 because there are some errors in the typesetting.

Point 6: Conclusion must specify the findings of the paper in numerical terms and clearly specify the future work.

Response 6: Thank you for underlining this deficiency. The conclusion was revised and modified according to the information showed in the work suggested by the reviewer, which not only specify the findings of the paper in numerical terms, but also clearly specify the future work.

Round 2

Reviewer 2 Report

Accept as it is. Congratulations

Author Response

we are gratetul for the sugestion and the comments. 

Reviewer 3 Report

Comments

Point 1: The authors have not added separate subsections on motivation, novelty, and contributions in the Introduction Section, which improves the readability of the paper.

Point 2: Still the authors have not added a comparison table in related works considering various parameters that clearly shows the difference.

Point 4: Errors in the equation references. It is printed in the manuscript.

 

 

Author Response

Response to Reviewer 3 Comments

 

Point 1: The authors have not added separate subsections on motivation, novelty, and contributions in the Introduction Section, which improves the readability of the paper.

 

Response 1: We deeply appreciate the reviewer’s suggestion. In order to improve the readability of the paper, we have added separate subsections on motivation, novelty, and contributions in the Introduction Section. We have introduced it in detail in the revised manuscript.

 

Point 2: Still the authors have not added a comparison table in related works considering various parameters that clearly shows the difference.

 

Response 2: We are very sorry for the above problem. In order to show the difference more clearly, we have added a comparison table in related works, as shown in Table 1.

TABLE I The related work comparisons

 

Paper

Technique

Work Goals

Region-based object detection of remote sensing images

2019 Dong et al. [11]

Faster R-CNN & sig NMS

Improve the problem of missing detection of small targets.

2021 Zheng et al. [12]

FAGNet

Improve object detection capabilities for small-scale remote sensing images.

2021 Yao et al. [13]

MFE-Net

Improve the performance of remote sensing image target detection

2022 Li et al. [14]

LF-CNN

Improve the accuracy of object detection in small sample remote sensing images.

Regression-based object detection of remote sensing images

2020 Courtrai et al. [18]

Wasserstein GAN

Improve the accuracy of small target detection in remote sensing images

2020 Yin et al. [19]

AF-SSD

Improve the accuracy of small-scale target detection in remote sensing images

2021 Yu et al. [20]

TWC-Net

Solve the problem of sparsity of image objects and complexity of interference

2021 Li et al. [21]

DSYOLOv3

Improve the detection effect of small objects in aerial images.

2021 Nong et al. [22]

Embedded-based Method

Reduce the number of parameters.

2022 Liu et al. [23]

NRT-YOLO

Solve the problem of detecting small and high resolution objects in remote sensing images.

2022 Fang et al. [24]

S2ANET-SR

Solve the problem that many small-scale objects in remote sensing images are difficult to detect.

This paper

HDAU-YOLO

Improve the accuracy of small target detection in remote sensing images

 

 

Point 4: Errors in the equation references. It is printed in the manuscript.

 

Response 4: We apologize for this problem in the manuscript. We have modified the equation references.

 

Round 3

Reviewer 3 Report

No comments.

Now the manuscript is updated and qualified for this journal publication. Keep doing good work.

Back to TopTop