Physically Based Neural Bidirectional Reflectance Distribution Function

TL;DR

This model accurately reconstructs real-world materials while uniquely enforcing physical properties for realistic BRDFs, specifically Helmholtz reciprocity via reparametrization and energy passivity via efficient analytical integration.

Abstract

We introduce the physically based neural bidirectional reflectance distribution function (PBNBRDF), a novel, continuous representation for material appearance based on neural fields. Our model accurately reconstructs real-world materials while uniquely enforcing physical properties for realistic BRDFs, specifically Helmholtz reciprocity via reparametrization and energy passivity via efficient analytical integration. We conduct a systematic analysis demonstrating the benefits of adhering to these physical laws on the visual quality of reconstructed materials. Additionally, we enhance the color accuracy of neural BRDFs by introducing chromaticity enforcement supervising the norms of RGB channels. Through both qualitative and quantitative experiments on multiple databases of measured real-world BRDFs, we show that adhering to these physical constraints enables neural fields to more faithfully and stably represent the original data and achieve higher rendering quality.

Figures (5)

• Figure 1: NBRDF sztrajman2021nbrdf violates Helmholtz reciprocity, leading to tangential discontinuities in the specular region (see diagonal discontinuity in the close-up inset of the zoomed-in image).
• Figure 2: Renderings of NBRDF sztrajman2021nbrdf and our PBNBRDF fitted to MERL BRDFs Matusik2003datadriven, along with the absolute error plots. Violating Helmholtz reciprocity creates tangential discontinuities in NBRDF renderings, whereas our PBNBRDF produces more natural and high-quality renderings without noticeable discontinuities.
• Figure 3: Renderings of NBRDF sztrajman2021nbrdf and our PBNBRDF fitted to the RGL dataset rgl2018 under the white furnace test heitz2014understanding. The non-passivity effect is exacerbated by NBRDFs, but is significantly reduced when using our PBNBRDF.
• Figure 4: Renderings of NBRDF and our PBNBRDF fitted to the RGL dataset rgl2018. NBRDF produces noticeable artifacts (fireflies) due to energy creation, whereas our PBNBRDF provides higher-quality renderings.
• Figure 5: Renderings of NBRDF sztrajman2021nbrdf and our PBNBRDF fitted to RGL BRDFs rgl2018, along with SSIM plots. PBNBRDF produces more realistic and natural renderings that are perceptually closer to the ground truth, as indicated by higher SSIM values (shallower color).