Data Driven SVBRDF Estimation Using Deep Embedded Clustering
Round 1
Reviewer 1 Report
You don't need to explain the first part explains something, the second part explains something and so on, but what you need to explain is what the problem is and what solutions you offer, what is the purpose of your research. there are still many theories that are submitted but are not supported by some previous research. In your article, the method you use has not been explained.
at the conclusion, describe the focus on the results of your research
Author Response
We thank you for all the valuable comments about the first draft. We, hereby, send you our point-by-point response.
Please check the attachment.
Best regards,
Yonghwi Kim
Author Response File: 
Author Response.pdf
Reviewer 2 Report
his paper presented an unsupervised deep-learning based method to estimate Spatially-Varying Bidirectional Reflectance Distribution Function (SVBRDFs) of a non-planar object. The paper is very well-written, the structure of it is very good, there are no grammar mistakes. The results are convincing.
Some minor remarks:
- SVBRDF was used in the abstract before the definition of the acronym.
- In the Introduction, authors wrote: "contrast to the aforementioned conventional methods, recent development of deep learning for SVBRDF..." without the main differences between conventional and a deep learning-based techniques.
- Authors explain a technique with: "It is specialized for an image classification problem where its kernel function measures the similarity between low-dimensional points and cluster centroids by an auxiliary target distribution based on the student t-distribution". This should be explained or a reference for the exact techniques should be provided.
- In Table 1., nc should be defined in the caption.
- In Table 1., results are hard to understand. Are the results dimensionless? It should be explained to ease the understanding.
- In Fig. 5. Axes labels are hard to understand. A short name would be useful.
- In Fig. 7., RMSE is dimensionless? - There are no comparisons with other techniques' accuracy. If there are such examples, comparisons should be provided.
Author Response
We thank you for all the valuable comments about the first draft. We, hereby, send you our point-by-point response.
Please check the attachment.
Best regards,
Yonghwi Kim
Author Response File: Author Response.pdf
Reviewer 3 Report
1. What is the novelty of the work.
2. What is the SVBRDF Estimation over deep leaning techniques.
Author Response
We thank you for all the valuable comments about the first draft. We, hereby, send you our point-by-point response.
Please check the attachment.
Best regards,
Yonghwi Kim
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
You don't need to explain the first part explains something, the second part explains something and so on, but what you need to explain is what the problem is and what solutions you offer, what is the purpose of your research.
there are still many theories that are submitted but are not supported by some previous research
In your article, the method you use has not been explained
Author Response
We appreciate your constructive comments about the revised manuscript. We attach our point-by-point response to your comments.
Please check the attachment.
Best regards,
Yonghwi Kim
Author Response File: Author Response.pdf
Reviewer 2 Report
The paper is significantly improved by the revision. Paper acceptance is recommended, after some minor fixes. Authors should comment on the outlook of the field, e.g. commenting on/ citing the optional use of thermal images: Molnar, A., Lovas, I. and Domozi, Z., 2021. Practical Application Possibilities for 3D Models Using Low-resolution Thermal Images. Acta Polytech. Hung, 18, pp.199-212. Also, their method's connection to other sensor modalities should also be discussed: Ngoc, T.T., Le Van Dai, C.M.T. and Thuyen, C.M., 2021. Support vector regression based on grid search method of hyperparameters for load forecasting. Acta Polytechnica Hungarica, 18(2), pp.143-158.
Author Response
We appreciate your comments about the revised manuscript. We attach our response to your comment about missing references.
Please check the attachment.
Best regards,
Yonghwi Kim
Author Response File: Author Response.pdf
