Letter Of Changes: We thank all the reviewers for their constructive comments.
---
### [Reviewer PJUU]: "The figures could be made more clear by improving the labels"
We have replaced ambiguously labels (e.g., Sun in Table 1, Figs. 3, 5, 6, and 9) with obvious ones
(e.g., previous method [26] and Ours (Fresnel model)).
We have added the MAPE inset images in Figs. 1 and 3.
We also have reduced the number of rows and columns in Fig. 3.
---
### [Reviewer bFXh]: "More careful analysis is needed for laika_ceilingpaint_18_gray material where our method is inferior to the previous method."
Since our method assumes that the material is well approximated with the microfacet model,
materials that deviate from this assumption are difficult to separate in our method.
This material is a waterborne paint material that violates this assumption and can exhibit volumetric effects (e.g. subsurface scattering).
It is interesting to separate such effects from the measured BRDF but we leave it for future work.
We have added the following sentence in the last paragraph of Sec. 5.1.
"It is conceivable that this waterborne paint material can violate our underlying assumption that the BRDF is modeled by the microfacet theory."
When submitting this paper, I forgot to register the first and the second authors in OpenReview
since this is the first time for me to use OpenReview.
I am sorry for this fault, but I would like to add the first and the second authors.
Best regards.
Keywords: BRDF, measured BRDF, diffuse-specular separation
TL;DR: This paper proposes a separation method of isotropic measured spectral BRDFs to handle goniochromatic effects and significantly improve the accuracy (up to 30 dB in PSNR).
Abstract: Measured BRDFs, which are acquired by measuring the reflectance of real-world materials, can reproduce the material appearance faithfully.
The measured reflectance is a mixture of reflection components with different properties, such as diffuse reflections and specular reflections.
Recent applications including light-probe rendering and denoising, however, often require a separate representation of BRDFs to apply
each component to the different pipelines for efficient rendering.
This paper proposes a separation method of isotropic measured spectral BRDFs to handle goniochromatic effects
based on the spectral microfacet BRDF model and diffraction microfacet BRDF model.
Experimental results show that our method can increase the PSNR up to about 30 dB compared with the previous method.
Supplementary Material: pdf
Submission Number: 11
Loading