A Renovated Framework of a Convolution Neural Network with Transformer for Detecting Surface Changes from High-Resolution Remote-Sensing Images

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript has a clear expression logic, sufficient theoretical interpretation and a sufficient amount of experiments. In general, the overall quality of this paper is excellent. However, from my perspective, there are some small problems and suggestions as follows：

1.       The contribution in the Introduction should be better described. Meanwhile, the motivation is relatively simple, it would be better if the author could give further exposition.

2.       Table 1 is spread across two pages, please adjust it to one page.

3.       Literature reviews rarely involve the optimization of spatial information, and authors should appropriately add introductions to relevant literature. such as “Deep Self-Representation Learning Framework for Detecting Surface Changes” and “Dynamic low-rank and sparse priors constrained deep autoencoders for Detecting Surface Changes”.

4.       In Section 3, to improve the readability of the article, some analysis of the underlying reasons should be added in the experiment results.

5.       Some advantageous parts in the detection maps can be marked with colored rectangular boxes, which can be more intuitive for the proposed method.

Comments on the Quality of English Language

English needs further improvement.

Author Response

Dear reviewer, 

We appreciate your valuable comments for improving this MS and please find attached file for the reponses. Many thanks.

Best,

Han Wang

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The authors show a new network structure with Transformer architecture and Residual network to detect changes in high-resolution remote sensing images.

The paper is clear and the method developed by the authors comes across well structured and organized. However, some aspects of the paper involving text formatting and the quality of some images require strong revision because they do not make clear the validity of the method. Regarding the text, the authors abuse boldface both in the writing and in the use of equations making it not easy to read; for example, if the use of a matrix is included in the equation then boldface is fine, otherwise it is strongly discouraged. The quality of figures 10-11-12 is poor and consequently need to be revised.

I also suggest introducing several literature references before equations 21-24.

Finally, more details about the type of satellite image used for experimentation should be provided (spatial resolution, satellite platform, etc.); consequently, also in the discussion part, the authors could discuss the importance of spatial resolution to apply the proposed algorithm.

Author Response

Dear reviewer, 

We appreciate your valuable comments for improving this MS and please find attached file for the reponses. Many thanks.

Best,

Han Wang

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have resolved these issues I raised.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have responded to all my comments and suggestions; in fact, the paper is greatly improved from the previous version. Consequently. the paper can be accepted in this version.