Next Article in Journal
Forest Height Mapping Using Feature Selection and Machine Learning by Integrating Multi-Source Satellite Data in Baoding City, North China
Previous Article in Journal
Spatiotemporally Continuous Reconstruction of Retrieved PM2.5 Data Using an Autogeoi-Stacking Model in the Beijing-Tianjin-Hebei Region, China
 
 
Review
Peer-Review Record

Few-Shot Object Detection in Remote Sensing Image Interpretation: Opportunities and Challenges

Remote Sens. 2022, 14(18), 4435; https://doi.org/10.3390/rs14184435
by Sixu Liu 1, Yanan You 1,*, Haozheng Su 1, Gang Meng 2, Wei Yang 3 and Fang Liu 1
Reviewer 1:
Reviewer 2:
Reviewer 3:
Remote Sens. 2022, 14(18), 4435; https://doi.org/10.3390/rs14184435
Submission received: 4 August 2022 / Revised: 28 August 2022 / Accepted: 2 September 2022 / Published: 6 September 2022
(This article belongs to the Section AI Remote Sensing)

Round 1

Reviewer 1 Report

In this paper, extensive papers on CV and RS related to Few-shot Object Detection research were analyzed, and the experimental results were presented. However, to improve this article's completeness, the following opinions are presented. 

1. Before explaining few-shot object detection, the general description of object detection and remote sensing object detection is too detailed. This is an opinion that I would like to develop the article by limiting to few-shot object detection in remote sensing. 

2. The contents of the previously published few-shot object detection in remote sensing papers and the contents of the experiment are described in detail in this paper. I would like the author to add a novel method proposed by the author and the experimental results and discussion on the proposed method.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

  Few-Shot Object Detection in Remote Sensing Image Interpretation: Opportunities and Challenges
In this article, the author first briefly introduces object detection task and their algorithms, in order to
better understand the basic detection frameworks followed by FSOD.
This manuscript is good, the writing style is also okay.

The paper is extremely in-depth and comprehensive. The author explains the literature in Figure 1 from 2018 to 2022. Literature cited in this paper. OD refers to object detection. NOT all FSOD works in the CV field are included because of their value in remote sensing image interpretation. Figure 3. Publication sources of literature in the field of remote sensing image interpretation (a) and computer vision (b). Also, the author shows A road map of deep learning object detection in Figure 6. The author also discusses The illustration of two-stage and one-stage object detectors in Figure 7.

1. This is a review paper or article? In my opinion, this is a review paper so please change it because the author state that it is an article paper in page 1.
2. Reduce the Figure, 20 figures in one manuscript is too much.
3. It is suggested to add some related equations and algorithms so this paper will be complete.
4. What are the limitations of Few-Shot Object Detection so far? Add a paragraph to explain.
5. Add future study directions related to this research. 

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

It is very important area for practical use of RS data on the ground. However, I have some comments and question about this study.

 

1.     What kind of RS data do you use in this study? It should be clearly mentioned about that because RS data type is key of your study including wavelength, spatial and temporal resolution.

2.     It seems that you just targeted stable material even if it is plane, ship, etc. If you target to check moving targets, what is effect in your research?

3.     What is processing performance and necessary computer resources with your proposed method?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

This paper summarizes the accuracy performance on experimental datasets to illustrate the achievements and shortcomings of the stated algorithms, and highlight the future opportunities and challenges of FSOD in RS image interpretation, in the hope of providing insights to the following research. The paper is very extensive and detailed. One of the things to correct is the language that must be improved scientifically. Another thing is related to the references, the authors must use "The authors of", "In" or "Last Name of author, et al." when they started a paragraph with a reference. Also, the discussion and comparison of the results with other studies must be presented.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have made a huge effort to complete and improve the old version which is now more close to be published in this prestigious journal. The opinions of reviewers were faithfully reflected.

Reviewer 4 Report

The manuscript is now improved and more readable, and it can be accepted. 

Back to TopTop