Image Inpainting with Bilateral Convolution

Round 1

Reviewer 1 Report

This paper is about image inpainting using a bilateral convolution strategy.

It seems like a mature manuscript. The authors have made an effort to explain all design decisions. The datasets used for experimentation are adequate and enough.

 

One minor comment is as follows:

 

- It is unclear why figure 1 in the introduction is a result figure as it contains “ours” results!

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Please see the attached file.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

In this paper, the authors customize a bilateral convolution network to reconstruct corrupted images, which stacks a series of BC blocks and MWA modules. With the BC operator, the BC block could consider the reliability of features based on the spatial location as well as the feature value simultaneously, which could adaptively handle diverse corrupted regions and efficiently preserve and propagate information from known regions to unknown regions. Furthermore, the MWA with multi-range window self-attention can model different range spatial dependencies among pixels. Quantitative and qualitative experiments on remote sensing datasets and typical image inpainting datasets, the results have demonstrated that the proposed network is robust to various scene images and generates more appropriate and consistent content than several state-of-the-art methods. The paper is interesting, well-written and can accepted after minor revision.

The authors should present loss curves.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

1.  1-  The presentation of methodology is unclear, need to rewrite accurately to understand its

2.  2-   Experimental results and discussion need to rewrite with add more analysis for results

4. 3- The presentation of results is unclear? Should be rewrite the experimental result with add more analysis  

5. 4- Rewrite the conclusion to include all findings. And explain what the limitations of this method.

6.      The authors should propose some ideas for future work in the “Conclusion” section.

7.      The English language of manuscript need to polish.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors have answered all of my concerns.