AI Paper Atlas

Research map / 3D vision / Depth and stereo

Materials for Masses: SVBRDF Acquisition with a Single Mobile Phone Image

Zhengqin Li, Kalyan Sunkavalli, Manmohan Chandraker

ECCV 2018 · 3D vision › Depth and stereo · approach: Neural light field and illumination imaging · 158 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add a material classifier whose outputs weight per-material BRDF predictions (ablation)encoder-decoder · SVBRDF testalbedo L2 error: 7.78 -> 7.58Improved
add a material classifier (ablation)encoder-decoder · SVBRDF testnormal L2 error: 1.55 -> 1.52Improved
add a material classifier (ablation)encoder-decoder · SVBRDF testroughness L2 error: 8.75 -> 8.55Improved
add DCRF refinement on top of the classifier (ablation)cls-pt · SVBRDF testalbedo L2 error: 7.58 -> 7.42Improved
add DCRF refinement on top of the classifier (ablation)cls-pt · SVBRDF testnormal L2 error: 1.52 -> 1.50Improved
add DCRF refinement on top of the classifier (ablation)cls-pt · SVBRDF testroughness L2 error: 8.55 -> 8.53Improved
train the classifier jointly with the BRDF network instead of as a pure classifier (ablation)material classifier · material classificationclassification error (%): 54.96 -> 73.65Improved
add a classifier but do not use its output to fuse BRDF predictions (ablation)encoder-decoder · SVBRDF testBRDF estimation errorNo clear effect
add a collocated point (flash) light to environment lighting at capture (ablation)cls network · SVBRDF testSVBRDF errorImproved
reduce relative flash intensity against environment light (ablation)full network · SVBRDF testSVBRDF errorHurt
use the proposed single-image network instead of the two-shot optimization methodSVBRDF reconstruction · stationary texture 192x192reconstruction time: 6 hours -> 2.4 secondsImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy