AI Paper Atlas

Research map / 3D vision / Face and shape

Shape Reconstruction Using Differentiable Projections and Deep Priors

Matheus Gadelha, Rui Wang, Subhransu Maji

ICCV 2019 · 3D vision › Face and shape · approach: Single-view 3D reconstruction · 38 citations (OpenAlex)

Read the paper

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
replace space carving with deep shape prior and view optimizationSilhouette reconstruction (8 views, perturbed viewpoints) · 12 shapes meanIoU: 0.55 -> 0.72Improved
use deep shape prior even with unperturbed viewpoints for the baselineSilhouette reconstruction · 12 shapes meanIoU: 0.61 -> 0.72Improved
perturb viewpoints for space carving (ablation)Space carving · 12 shapes meanIoU: 0.61 -> 0.55Hurt
use deep shape prior instead of space carving on planeSilhouette reconstruction · planeIoU: 0.35 vs 0.49Hurt
use deep shape prior instead of space carving on lampSilhouette reconstruction · lampIoU: 0.62 vs 0.26Improved
use deep image prior instead of iterative BM3DTomography (n=30, sigma=1) image 1 · Shepp-Logan-type imageSSIM: 0.701 -> 0.967Improved
use deep image prior instead of iterative BM3DTomography (n=30, sigma=1) image 1 · Shepp-Logan-type imagePSNR: 26.0 -> 31.7Improved
use deep image prior instead of iterative BM3DTomography (n=30, sigma=1) image 2 · BrainWeb slicePSNR: 26.2 -> 26.8Improved
use deep image prior instead of iterative BM3DTomography (n=30, sigma=1) image 3 · BrainWeb sliceSSIM: 0.718 -> 0.834Improved
add TV prior to FBP reconstructionTomography (n=30, sigma=1) · image 1PSNR: 15.3 -> 18.6Improved
increase number of views for depth reconstruction (4 to 32) at high noise (ablation)Depth-image reconstruction · sigma=0.1reconstruction quality (qualitative)Improved

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