Image Retrieval for Local Architectural Heritage Recommendation Based on Deep Hashing
Round 1
Reviewer 1 Report
Review formal aspects of writing the paper, i.e.:
58 china
59 japan
In the text, the scope of the analysis of the images must be recognized, as a part of the formal fact of heritage assets, which is a relevant factor but represents only a part of something much more complex.
Local architecture heritage images are related to certain classes defined in CAH10. It must be justified that all these classes are significant and consistent with respect to the historical architecture protection catalogs of cultural institutions and the legal principles on which said catalogs are based. It must also be justified that they are consistent with the way in which the historical heritage is collected in the documents of the urban plans that include historical areas or historical elements. And it must be justified how the method used, which indicates that it should be opened to more categories, can help in the revisions of heritage and urban catalogs, with the use of GIS, among other instruments.
It is recommended to comment in greater depth on the exceptions of the confusing classes identified in figure 4.
Author Response
Thanks for your valuable comments. We carefully revised the manuscript.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
In this paper, the authors proposed a hashing-based image retrieval method for local architectural heritage recommendations, which aims at a real-world scene where only a few data is available for local architectural heritage. It is meaningful for the promotion and protection of local architectural heritage.
The paper is nicely structured and paced. I enjoyed reading it and gained a lot of new knowledge. The results are well presented and necessary comparisons between many different methods are provided. Appropriate conclusions are made based on the acquired values.
More importantly, the article provides open-source code, which strengthens the reliability of the work and helps interested researchers follow and pursue related research.
However, some minor typos and inconsistencies are required to be amended. Some examples of that are as follows:
The same loss function should be used for the letter "L" in Eqs. 2, 5 and 6 in the same form, and a multiplication sign seems to be missing in Eq. 6.
Line 232: I would like to confirm whether the initial learning rate here should be 5e-4?
Line 189/301/329/349: the "s" in the word "section" should be capitalised like the form in Line 226.
Line 372: ... more state-of-art technologies. -> ... more the state-of-the-art technologies.
There are a few mistakes like that, so please double-check the entire paper.
Appropriate references need to be added in the first paragraph of the introduction.
I suggest this paper for publication after minor revision.
Thank you
Author Response
Thanks for your valuable comments. We carefully revised the manuscript.
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper presents a study on the use of Image Retrieval for Local Architectural Heritage Recommendation Based on Deep Hashing. The introduction is not clear. Some part of the background creation could be inserted in the introduction, to have a more general outline on this topic. I suggest o also to ref to the paper https://doi.org/10.1016/j.enbuild.2022.112029 that have a section dedicated to the literature review of your topic. This could help you to re-ordinate the ideas at the basis of the study compared to other study. Also you can find the novelty of you paper. There are also several English mistakes, thus a mother tongue revision is needed. The description of the tool is clear, but comparisons with others examples could be useful for understanding its scientific value. Conclusions must be rewritten.
Author Response
Thanks for your valuable comments. We carefully revised the manuscript.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
The author addressed all my comments in a positive way, adding new overview on recent researches, exposing the novelty and changing the conclusions.Thus, I can accept the paper in the present version.
