Self-Supervised Monocular Depth Estimation Based on Channel Attention

Round 1

Reviewer 1 Report

The paper looks good to me. 

Author Response

Dear Editors:
We appreciate the detailed comments from the reviewers. We have responded to the comments on for the reviewers. The changes in the manuscript resulting from the comments were identified in the answers as well. As you will find out, this paper has been carefully revised to address the comments from the reviewers. 
We thank you for your and the reviewer’s time in helping us to make this a better manuscript. As you can see from the revised manuscript, we did our best to address the comments from the reviewers. We also made extensive revision to the manuscript based on the reviews. As a result, the quality of the manuscript was dramatically improved with the help from our reviewers. Should you have any other questions, please feel free to let us know.
Best regards,
Dr. Bo Tao

Author Response File: Author Response.docx

Reviewer 2 Report

Excellent presentation of improved new network structure based on a self-supervised monocular depth estimation model, with encouraging result and superiority outlined.

Author Response

Dear Editors:
We appreciate the detailed comments from the reviewers. We have responded to the comments on for the reviewers. The changes in the manuscript resulting from the comments were identified in the answers as well. As you will find out, this paper has been carefully revised to address the comments from the reviewers. 
We thank you for your and the reviewer’s time in helping us to make this a better manuscript. As you can see from the revised manuscript, we did our best to address the comments from the reviewers. We also made extensive revision to the manuscript based on the reviews. As a result, the quality of the manuscript was dramatically improved with the help from our reviewers. Should you have any other questions, please feel free to let us know.
Best regards,
Dr. Bo Tao

Author Response File: Author Response.docx

Reviewer 3 Report

This paper presents a depth prediction network using monocular images. It adopts two channel attention modules, structure perception module and detail emphasizing module. The approach for the implementation are described. Some experimental results and comparison with other method are provided. There are several issues the authors have to further address in the revised manuscript. First, the roles and usages of the two adopted modules are not very clear. Although some detailed descriptions and formation are provided, it is not shown how they are contribution to the depth prediction network. Second, it seems that the backbone is U-Net, and PoseNet is also used. U-Net is commonly known for image segmentation, not depth estimation. The authors need to explain more precisely how it is used for depth prediction. Also, why pose network needs to be used? Third, the introduction and literature survey are not comprehensive. There are many works related to depth prediction for the outdoor scenes. More recent work such as

Wang, H.M., Lin, H.Y. and Chang, C.C., 2021. Object Detection and Depth Estimation Approach Based on Deep Convolutional Neural Networks. Sensors, 21(14), p.4755.

Ming, Y., Meng, X., Fan, C. and Yu, H., 2021. Deep learning for monocular depth estimation: A review. Neurocomputing, 438, pp.14-33. should be referenced and discussed. Fourth, the experiments are carried out and comparison with other methods are given. However, the discussion is missing. It is suggested to show more detailed experiment information, as well as providing the source code of this work for evaluation.

 

Author Response

Dear Editors:
We appreciate the detailed comments from the reviewers. We have responded to the comments on for the reviewers. The changes in the manuscript resulting from the comments were identified in the answers as well. As you will find out, this paper has been carefully revised to address the comments from the reviewers. 
We thank you for your and the reviewer’s time in helping us to make this a better manuscript. As you can see from the revised manuscript, we did our best to address the comments from the reviewers. We also made extensive revision to the manuscript based on the reviews. As a result, the quality of the manuscript was dramatically improved with the help from our reviewers. Should you have any other questions, please feel free to let us know.
Best regards,
Dr. Bo Tao

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

This revision has addressed most of the reviewer's concerns. It can be accepted with further grammar and typo checks.

Author Response

Dear Editors:
We appreciate the detailed comments from the reviewers. We have responded to the comments on for the reviewers. The changes in the manuscript resulting from the comments were identified in the answers as well. As you will find out, this paper has been carefully revised to address the comments from the reviewers. 
We thank you for your and the reviewer’s time in helping us to make this a better manuscript. As you can see from the revised manuscript, we did our best to address the comments from the reviewers. We also made extensive revision to the manuscript based on the reviews. As a result, the quality of the manuscript was dramatically improved with the help from our reviewers. Should you have any other questions, please feel free to let us know.
Best regards,
Dr. Bo Tao

Author Response File: Author Response.pdf