AI Paper Atlas

Research map / 3D vision / Depth and stereo

OPAL: Occlusion Pattern Aware Loss for Unsupervised Light Field Disparity Estimation

Peng Li, Jiayin Zhao, Jingyao Wu, Chao Deng, Yuqi Han, Haoqian Wang, Tao Yu

TPAMI 2023 · 3D vision › Depth and stereo · approach: Depth from light field and defocus · 38 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 OPAL occlusion-pattern loss (ablation)OPENet-fast · 4D LF Benchmark (4-scene average, ablation)MSE x100: 3.828 -> 1.978Improved
simplify occlusion patterns with beta=4 instead of beta=1 (ablation)OPENet-fast · 4D LF Benchmark (4-scene average, ablation)MSE x100: 1.978 -> 1.908Improved
use non-simplified patterns beta=1 instead of beta=4 (ablation)OPENet-fast · 4D LF Benchmark (Sideboard)MSE x100: 1.141 -> 1.000Improved
add refinement module (ablation)OPENet · 4D LF Benchmark (4-scene average, ablation)MSE x100: 1.820 -> 1.672Improved
add EPI-Transformer (ablation)OPENet · 4D LF Benchmark (4-scene average, ablation)MSE x100: 1.808 -> 1.672Improved
add raw-disparity loss term (ablation)OPENet · 4D LF Benchmark (4-scene average, ablation)MSE x100: 1.755 -> 1.672Improved
add smoothness loss (ablation)OPENet · 4D LF Benchmark (4-scene average, ablation)MSE x100: 2.482 -> 1.672Improved
use unsupervised OPENet instead of supervised LFattNet across datasetsOPENet · 4D LF Benchmark + HCI Blender (cross-dataset average)MSE x100: 2.164 -> 1.127Improved
use OPENet instead of OPENet-fast (EPI-Transformer, refinement, four directions)OPENet · 4D LF Benchmark + HCI Blender (cross-dataset average)MSE x100: 1.315 -> 1.127Improved

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