Reducing System Load of Effective Video Using a Network Model

Round 1

Reviewer 1 Report

The paper proposes a new method to reduce the amount of  video data without decreasing the quality with a machine learning method. For increasing the readability, I recommend the following modifications.

1) In the last of Section 1, the paper needs to clarify the contribution of the study. In the last paragraph of Section 1, the paper presents an overview of the proposed approach, but it is not clear the novelty of the approach.

2) The weakest point of the paper is that the paper shows only quantitative results. There is no discussions to interpret the data extracted from the simulation. The paper needs to include more discussions why the data shown in Section 4.3 are significant.

3) What is SRCNN ? Also, it is better to explain why the study choose SRGAN and SRCNN.

4) Section 4.3 is too short and ambiguous. Especially, videos used in the study are not clear. The section should show what kind of games or streaming videos were used. If not, it is hard to believe the results.

Author Response

1) In the last of Section 1, the paper needs to clarify the contribution of the study. In the last paragraph of Section 1, the paper presents an overview of the proposed approach, but it is not clear the novelty of the approach.

Response 1: Please see the attachment. (Page 2)

2) The weakest point of the paper is that the paper shows only quantitative results. There is no discussions to interpret the data extracted from the simulation. The paper needs to include more discussions why the data shown in Section 4.3 are significant.

Response 2: Please see the attachment. (Page 13)

3) What is SRCNN ? Also, it is better to explain why the study choose SRGAN and SRCNN.

Response 3: Please see the attachment. (Page 6, 11)

4) Section 4.3 is too short and ambiguous. Especially, videos used in the study are not clear. The section should show what kind of games or streaming videos were used. If not, it is hard to believe the results.

Response 4: Please see the attachment. (Page 13)

Author Response File: Author Response.docx

Reviewer 2 Report

The paper has proposed a method in order to reduce the system load without deteriorating QoE. In particular, authors propose to tackle some issues related to real-time streaming services such as network congestion, delay, packet loss, etc that can cause performance degradation.

 

There are some changes needed in the paper which are as follows:

• To highlight the contribution in the paper there should be a subsection as contribution where authors need to define.
• I assume that some of the formulas used in the work are not original. Then please associate relevant bibliographic references to each of them.
• Some references are needed in the introduction section especially for DAIN, SSAS and SRGAN techniques. At this level authors should clarify what is the novelty and contribution as compared to the state of the art. Please detail the individual steps in more detail.
• Fix some grammatical errors such as in line 51 (“First, the image was segmentation….) and in line 297 (titles of fig 17 and 18 are the same. They should be different), etc.
• Evaluation section is not well presented. Metrics are not well defined. Comparative performances are not clear at all. This section should be clear and results well explained because there are a lot of unexplained tables.
• There are only self-compared ablation experiments. No comparison results with the state of the arts are performed.
• Please describe the available image dataset in more detail.

Author Response

Some references are needed in the introduction section especially for DAIN, SSAS and SRGAN techniques. At this level authors should clarify what is the novelty and contribution as compared to the state of the art. Please detail the individual steps in more detail.

Response 1: Please see the attachment.(PAGE 1,6 - add to 2.5 SRCNN)

Fix some grammatical errors such as in line 51 (“First, the image was segmentation….) and in line 297 (titles of fig 17 and 18 are the same. They should be different), etc.

Response 2: Please see the attachment.(Add figure and change figure title.)

Evaluation section is not well presented. Metrics are not well defined. Comparative performances are not clear at all. This section should be clear and results well explained because there are a lot of unexplained tables.

There are only self-compared ablation experiments. No comparison results with the state of the arts are performed.
Please describe the available image dataset in more detail.

Response 3: Please see the attachment.(PAGE 14 - 4.3 Evaluation)

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The author's response contains only the revised version of the paper. There is no answers for my comments so I could not understand how the paper is changed according to my comments. The response should explain how the revised version was changed according to the comments.

Author Response

Each answer to the comment is marked in red and a note is attached.

1) In the last of Section 1, the paper needs to clarify the contribution of the study. In the last paragraph of Section 1, the paper presents an overview of the proposed approach, but it is not clear the novelty of the approach.

Response 1: I wrote a detailed step-by-step description of the approach proposed in the last paragraph of Section 1. (Page2)

2) The weakest point of the paper is that the paper shows only quantitative results. There is no discussions to interpret the data extracted from the simulation. The paper needs to include more discussions why the data shown in Section 4.3 are significant.

Response 2: Images used for the data indicated in Section 4.3 were included, and explanations of the tables were described accordingly. (Page 14)

3) What is SRCNN ? Also, it is better to explain why the study choose SRGAN and SRCNN.

Response 3: Section 2.5 with attached data on SRCNN and Added Section 3.4, which describes the reasons for choosing SRGAN and SRCNN. (Section 2.5 = Page6, Section 3.4 = Page11)

4) Section 4.3 is too short and ambiguous. Especially, videos used in the study are not clear. The section should show what kind of games or streaming videos were used. If not, it is hard to believe the results.

Answer 4: Comments 2 and 4 were written in Section 4.3. (Page 14)

Author Response File: Author Response.docx

Reviewer 2 Report

The authors addressed the overall of my comments. This paper may be accepted in its current form. No more comments. Thank you.

Author Response

Thanks for your comments on this paper.

Round 3

Reviewer 1 Report

I think that the revised paper reflects my comments.

Author Response

Thanks for your comments on this paper.