Training of Convolutional Neural Networks for Image Classification with Fully Decoupled Extended Kalman Filter

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In this work, the authors carried out a study for training of CNN in the context of image classification. The topic is interesting and worth investigation, however I have the following concerns:

1- Major: The paper in its current form is read like applied research as it applies a well-known concept to a well-known problem. So, what is the main scientific contribution is not clear and seems quite narrow. 

2- Minor: The introduction section is quite long due to the inclusion of related works. I recommend simplifying the introduction section in a way that provides the main context, the problem and motivation of the work, the proposed solution, and the main contribution. 

3- Minor: The related works that have been integrated in the introduction can be moved to the section 4 related works and at the end you can highlight the use of three algorithms as a foundation for your work. 

4- Minor: In the introduction section, the last paragraph on page 2 has been repeated at the beginning of page 3 as well. so, pay attention to this type of error. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The authors propose an approach to the EKF for training CNNs in image classification concepts. Moreover, some comparative analyses are performed against the Adam optimizer, the sKAdam algorithm, and reduced EKF. The manuscript is written well, and the flow of this work is clear. However, some comments on the manuscript could be considered as follows:
•
The authors should scientifically focus on the research gap(s) and add them in the Abstract and Introduction. These research gap(s) could be only two or a maximum of three targets the authors already achieved.
•
All equations should be clarified by their parameters’ definitions.
•
The authors mentioned the existing works well but the novelty of their work MUST be highlighted clearly.
•
The main drawback of the manuscript is that the authors used very limited images for their experiments (Fig. 3).
•
The quality of graphs is poor (Figs. 4, 5 and 6). They have to update their graphs with high-resolution images.
•
Some related references could be added to this manuscript. The authors should modify the manuscript regarding the following scholars to enrich the literature review.
o Chen, B., Zhang, L., Chen, H., Liang, K. and Chen, X., 2021. A novel extended Kalman filter with support vector machine based method for the automatic diagnosis and segmentation of brain tumors. Computer Methods and Programs in Biomedicine, 200, p.105797.
o Lu, A. and Honarvar Shakibaei Asli, B., 2023. Seismic Image Identification and Detection Based on Tchebichef Moment Invariant. Electronics, 12(17), p.3692.
o
Movaghati, S., Moghaddamjoo, A. and Tavakoli, A., 2010. Road extraction from satellite images using particle filtering and extended Kalman filtering. IEEE Transactions on geoscience and remote sensing, 48(7), pp.2807-2817.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

A general edit is needed.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have addressed my concerns and know the main contributions are more clear.