AI Paper Atlas

Research map / 3D vision / Face and shape

Learning to Reconstruct High-quality 3D Shapes with Cascaded Fully Convolutional Networks

Yan-Pei Cao, Zheng-Ning Liu, Zheng-Fei Kuang, Leif Kobbelt, Shi-Min Hu

ECCV 2018 · 3D vision › Face and shape · approach: Learned mesh and geometric representations · 32 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 visual hull input branches (ablation)3D-CFCN (2 views) · Human10 reconstructionrelative Hausdorff RMS distance: 0.0035 -> 0.0031Improved
add visual hull input branches (ablation)3D-CFCN (2 views) · ModelNet40 reconstructionrelative Hausdorff RMS distance: 0.0032 -> 0.0019Improved
fuse 4 viewpoints instead of 2 (ablation)3D-CFCN · Human10 reconstructionrelative Hausdorff RMS distance: 0.0035 -> 0.0021Improved
fuse 4 viewpoints instead of 2 (ablation)3D-CFCN · ModelNet40 reconstructionrelative Hausdorff RMS distance: 0.0032 -> 0.0010Improved
use 3D-CFCN instead of OctNetFusion3D-CFCN vs OctNetFusion · Human10 reconstructionrelative Hausdorff RMS distance: 0.0040 -> 0.0035Improved
use 3D-CFCN instead of OctNetFusion3D-CFCN vs OctNetFusion · ModelNet40 reconstructionrelative Hausdorff RMS distance: 0.0035 -> 0.0032No clear effect
use 3D-CFCN instead of 3D-EPN3D-CFCN vs 3D-EPN · Human10 reconstructionrelative Hausdorff RMS distance: 0.0263 -> 0.0035Improved
use 3D-CFCN instead of PSR3D-CFCN vs PSR · ModelNet40 reconstructionrelative Hausdorff RMS distance: 0.0620 -> 0.0032Improved
add second refinement stage (ablation)3D-CFCN · Human10 reconstructionvisual qualityImproved
add second refinement stage (ablation)3D-CFCN · ModelNet40 reconstructionvisual qualityNo clear effect
use 3D-CFCN decoder instead of OGN decoder on same features3D-CFCN vs OGN · Human10 occupancy maps (512^3)visual fidelityImproved
perturb camera poses by 2.5% to 10% at test time3D-CFCN · calibration error testvisual qualityMixed

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