Optimizing Generative Adversarial Network (GAN) Models for Non-Pneumatic Tire Design
Round 1
Reviewer 1 Report
The current work looks at generating designs of pneumatic tires using various GAN architectures. The scientific explanations are sound, with conclusions supporting the results. Some additional corrections are needed for publication:
1.) Intro: Introduction contains no information about pneumatic tires. This is the main area of application, so details are necessary.
2.) Extra space in Line 38.
3.) Line 69: some literature background is necessary to understand what kind of shapes of such tires are practical. That would improve genuine-ness of the current study use case
4.) line 74: What was the original size of the images ? how are the pixel sizes improved ? extrapolation ? is there evidence to show that this method doesn't affect final results ? A visual example would help here.
5.) Table 2: Inconsistently made table, unknown language on the right top
6.) How do the errors obtained in the current work compare with those in lit ? Maybe the values will give some insight, as compared to people and pokemon, it might be easier to capture objects with more regular geometry ?
English quality seems fine with minor grammar and spelling checks needed
Author Response
Intro: Introduction contains no information about pneumatic tires. This is the main area of application, so details are necessary.
- -> We have added information about pneumatic and non-pneumatic tires from lines 31 through 42.
Extra space in Line 38.
- ->We removed the extra space in line 38
Line 69: some literature background is necessary to understand what kind of shapes of such tires are practical. That would improve genuine-ness of the current study use case
- ->Literature backgrounds have been added from line 72 to 76.
- ->Added a new image (Figure 1. Original images from Korea Automotive Technology Institute) for an example of shapes.
Line 74: What was the original size of the images? how are the pixel sizes improved ? extrapolation ? is there evidence to show that this method doesn't affect final results ? A visual example would help here
- ->We have added original image information such as pixel and resize information from line 90 to 92.
Table 2: Inconsistently made table, unknown language on the right top
- ->We fixed the table2.
How do the errors obtained in the current work compare with those in lit ? Maybe the values will give some insight, as compared to people and pokemon, it might be easier to capture objects with more regular geometry ?
- -> Explained the errors received in line 417. Added the reason why shape and pattern are different between humans and Pokémon in line 417 ~ 419.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Dear Authors,
The paper has been well introduced, and the methods are adequately described.
I have some comments about the paper:
1. Could you provide a more detailed description of the application scenario? Additionally, it would be beneficial if you could explain how these methods could enhance the creativity of tire designers.
2. Please translate the title of the last column in Table 2 into English.
There are no specific issues with the English language to report, except for some formatting errors and typos. Careful proofreading is all that's needed.
Author Response
- Could you provide a more detailed description of the application scenario? Additionally, it would be beneficial if you could explain how these methods could enhance the creativity of tire designers.
-> Added information about pneumatic and non-pneumatic tires in line 31 ~ 42
-> Provided example picture in new Figure 1.
- Please translate the title of the last column in Table 2 into English.
-> Fixed. Thanks for your feedback.
Author Response File: Author Response.docx
