Next Article in Journal
Multi-Hazard Exposure Mapping Using Machine Learning for the State of Salzburg, Austria
Previous Article in Journal
An Approach to Minimize Atmospheric Correction Error and Improve Physics-Based Satellite-Derived Bathymetry in a Coastal Environment
 
 
Article
Peer-Review Record

A Novel Deep Forest-Based Active Transfer Learning Method for PolSAR Images

Remote Sens. 2020, 12(17), 2755; https://doi.org/10.3390/rs12172755
by Xingli Qin 1, Jie Yang 1, Lingli Zhao 2,*, Pingxiang Li 1 and Kaimin Sun 1
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Remote Sens. 2020, 12(17), 2755; https://doi.org/10.3390/rs12172755
Submission received: 22 July 2020 / Revised: 18 August 2020 / Accepted: 21 August 2020 / Published: 25 August 2020
(This article belongs to the Section Remote Sensing Image Processing)

Round 1

Reviewer 1 Report

Dear Authors:

Your proposal about PolSAR images offers interesting results; however, there are some issues that in my opinion need to be addressed in order of having a finished and suitable for publication paper. 

  1. Please, provide all the parameters used for the kernel of the SVM weak classifier.
  2. Did you tried to modify those parameters in order of improving the results, try to add more information in your experiments section. 
  3. In page 19, you say "... the proposed method is significantly better than that the other methods..."; there are some parametric or no parametric test that could help you to assure that the results are really significant, try to apply at least one test to confirm your statement, or the results could be described as equivalent. 
  4. Try to add at least another experiment to have some extra evidence of the good results of your work. 
  5. Try to summarize the results in a single table, highlighting the best performances and add the execution time, to assess the efficency of your method. 

I hope you find those recommendations useful to improve the final version of your work. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Novel technique and deserve to be published!

However the language needs to be improved!  

 

 

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper proposed a robust method of active transfer learning using the deep forest. The experiment showed a higher performance of transfer learning than past methods. As the authors solidly described the method and examined the results in sufficient aspects, I think the paper is almost ready for publication. Therefore, I would recommend this for publication though I would like to ask the authors some minor changes as below.

a. Abstract - please cite some quantified indicators or numbers from the results, such as Figure 13 and Figure 14, like "proposed method achieved 83%-88% whereas the accuracy of the past methods were 63%-83%."
b. 3. Experiments - please add descriptions of ground truth data preparation. Was the data prepared by visual interpretation of the satellite images?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Dear authors:

I have finished the review of your paper. I have no more recommendations or concerns about your work.

Back to TopTop