AI Paper Atlas

Research map / 3D vision / Face and shape

Geometric Correspondence Fields: Learned Differentiable Rendering for 3D Pose Refinement in the Wild

Alexander Grabner, Yaming Wang, Peizhao Zhang, Peihong Guo, Tong Xiao, Peter Vajda, Peter M. Roth, Vincent Lepetit

ECCV 2020 · 3D vision › Face and shape · approach: Articulated and non-rigid shape learning · 8 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
use a larger receptive field (7x7 kernels) (ablation)correspondence branch · Pix3D MedErr_R,tMedErr_R,t (x1e-2): 1.75 -> 1.34Improved
use more layers (4 instead of 2) in correspondence branch (ablation)correspondence branch · Pix3D MedErr_R,tMedErr_R,t (x1e-2): 1.58 -> 1.34Improved
use deeper mapping branches instead of a single Conv-BN-ReLU block (ablation)mapping branch · Pix3D MedErr_R,tMedErr_R,t (x1e-2): 1.55 -> 1.34Improved
add geometric attention (ablation)Our Refinement · Pix3D MedErr_R,tMedErr_R,t (x1e-2): 1.45 -> 1.34Improved
concatenate depth, normal and object coordinate renderings instead of mask only (ablation)Our Refinement · Pix3D MedErr_R,tMedErr_R,t (x1e-2): 1.44 -> 1.34Improved
concatenate three renderings instead of depth only (ablation)Our Refinement · Pix3D MedErr_R,tMedErr_R,t (x1e-2): 1.38 -> 1.34Improved
concatenate three renderings instead of normal only (ablation)Our Refinement · Pix3D MedErr_R,tMedErr_R,t (x1e-2): 1.36 -> 1.34Improved
replace Mask Refinement with learned correspondence-field refinementMask Refinement · Pix3D (GT models) MedErr_R,tMedErr_R,t (x1e-2): 2.96 -> 1.34Improved
replace Image Refinement with learned correspondence-field refinementImage Refinement · Pix3D (GT models) MedErr_R,tMedErr_R,t (x1e-2): 4.31 -> 1.34Improved
use retrieved 3D models instead of ground-truth modelsOur Refinement · Pix3D MedErr_R,tMedErr_R,t (x1e-2): 1.34 -> 2.14Hurt
add refinement with retrieval on top of feed-forward pose baselineBaseline · Pix3D MedErr_R,tMedErr_R,t (x1e-2): 4.76 -> 2.14Improved
run Image Refinement for 1000 iterations instead of 20Image Refinement · Pix3DMedErr: 20 iterations best vs 1000Hurt

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