AI Paper Atlas

Research map / Point clouds / Shape and surface completion

Deep Implicit Surface Point Prediction Networks

Rahul Venkatesh, Tejan Karmali, Sarthak Sharma, Aurobrata Ghosh, R. Venkatesh Babu, László A. Jeni, Maneesh Singh

ICCV 2021 · Point clouds › Shape and surface completion · approach: Distance-function surface reconstruction · 4 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 projection step to sphere tracing (ablation)CSP sphere tracing · ShapeNetDepth error (MAE): 0.016 -> 0.014Improved
use Jacobian normals instead of forward-mode normals at alpha=0CSP · ShapeNetNormal map similarity: 0.620 -> 0.913Improved
step back along ray by alpha=0.005 for forward-mode normals (ablation)CSP (fwd.) · ShapeNetNormal map similarity: 0.620 -> 0.912Improved
increase step-back alpha from 0.005 to 0.05 for Jacobian normals (ablation)CSP (jac.) · ShapeNetNormal map similarity: 0.920 -> 0.871Hurt
use CSP forward normals instead of NDF gradient normalsCSP (fwd.) · Rendering 512x512 imageRendering time: 0.063s -> 0.003sImproved
use CSP instead of NDF on unprocessed shapesCSP · ShapeNet unprocessedNormal cosine similarity: 0.86 -> 0.92Improved
use CSP instead of NDF on unprocessed shapesCSP · ShapeNet unprocessedChamfer-L2 (x10^4): 1.73 -> 1.28Improved
use CSP instead of SAL on unprocessed shapesCSP · ShapeNet unprocessedDepth error: 0.025 -> 0.014Improved
use CSP instead of NDF on closed shapesCSP · ShapeNet closed shapesChamfer-L2 (x10^4): 0.05 -> 0.1Hurt
use per-class fine-tuned CSP instead of DeepSDFCSP · ShapeNet SofasChamfer-L2 (x10^4): 3.29 -> 0.32Improved
use coarse-to-fine meshing instead of NDF point cloud plus BPACSP meshing · representative ShapeNet exampleMeshing time: 104s -> 2.50sImproved
use coarse-to-fine meshing instead of dense evaluationCSP meshing · representative ShapeNet exampleFunction evaluations: 12M -> 636kImproved

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