Entropy-Mediated Decision Fusion for Remotely Sensed Image Classification

Round 1

Reviewer 1 Report

The suggested revisions have been applied.

Reviewer 2 Report

The method is new and interesting, properly argumented and proved by relevant cases.

I just suggest some small corrections (see comments in attached pdf), and point out minor grammar issues (empty comments in the pdf)

Comments for author File: Comments.pdf

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

The authors propose a novel HSI classification method by decision fusion. The proposed method is interesting, but not novel enough in my view. Actually it is more like a regular multi-task learning problem. Some minor suggestions are as follows.

The main problems should lie in the experiment part. first, more comparison should be made, with more state of the art methods, as I think the achieved accuracies on Indian Pines seem not good enough. Second, more datasets should be tested to verify the effectiveness of the proposed framework. 

Some hyper parameters need to be made clear, like how you balance the soft decision and hard decision, how to choose the weight. All the hyper parameters in the model need to be made clear in the experiment part.

For the 10% training samples, how you choose them? randomly? and did you use introduce cross fold to complete the classification?

The authors may refer to the following papers and cite them.

1. Active learning incorporated deep transfer learning for hyperspectral image classification, IEEE journal of selected topics in applied earth observation and remote sensing, 2018.

2. Structure preserving transfer learning for unsupervised hyperspectral image classification, IEEE geoscience and remote sensing letter, 2018.

Reviewer 2 Report

The authors have investigated into image processing by the concept of entropy. The mathematical style is correct. Some improvements could be made with regards to the
references. In fact, the authors should include the references below (at least).

1) S. Li , X. Kang , L. Fang , J. Hu , H. Yin , Pixel-level image fusion: a survey of the state of the art, Inf. Fusion 33 (January) (2017) 100–112.

2) Y. Liu , S. Liu , Z. Wang , A general framework for image fusion based on multi-s- cale transform and sparse representation, Inf. Fusion 24 (July) (2015) 147–164.

3) R.C. Guido , A note on a practical relationship between filters coefficients and the scaling and wavelet functions of the discrete wavelet transform, Appl. Math. Lett. 24 (n.7) (2011) 1257–1259.

4) R.C. Guido , S. Barbon Jr , L.S. Vieira , et al. , Introduction to the discrete shapelet transform and a new paradigm: joint time-frequency-shape analysis, in: Proc. IEEE International Symposium on Circuits and Systems (IEEE ISCAS 2008), Seattle, WA, USA, vol. 1, 2008, pp. 2893–2896 .

5) E. Guariglia. Entropy and Fractal Antennas, Entropy (2016), 18(3), 84.

Both references are justified from the increasing importance that both the wavelet analysis and the concept of entropy cover in image processing.

Moreover, I strongly recommend an additional English review.