AI Paper Atlas

Research map / 3D vision / Face and shape

Deep Learning on Object-Centric 3D Neural Fields

Pierluigi Zama Ramirez, Luca De Luigi, Daniele Sirocchi, Adriano Cardace, Riccardo Spezialetti, Francesco Ballerini, Samuele Salti, Luigi Di Stefano

TPAMI 2024 · 3D vision › Face and shape · approach: Learned mesh and geometric representations

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 ReLU instead of SIREN activation in input NFs (ablation)nf2vec · ModelNet40 classification (UDF)accuracy (%): 87.0 -> 88.1Improved
initialize all NFs with the same random vector (ablation)nf2vec · NF weight alignmentconvergence / linear mode connectivityImproved
replace shared-network NFs (Functa) with individual NFsNF reconstruction · Manifold40 mesh reconstructionChamfer distance (mm): 2.85 -> 0.26Improved
replace shared-network NFs (DeepSDF) with individual NFsNF reconstruction · Manifold40 mesh reconstructionF-score (%): 25.1 -> 69.7Improved
use the row-wise shared encoder instead of a flattened-weights MLP encodernf2vec encoder · 4-layer 512-unit NFencoder parameters: ~800M -> ~3MImproved
classify NeRF embeddings instead of rendered single viewResNet50 · ShapeNetRender NeRF classificationaccuracy (%): 86.88 -> 87.28No clear effect
classify NeRF embeddings instead of nine rendered viewsResNet50 multi-view · ShapeNetRender NeRF classificationaccuracy (%): 93.28 -> 87.28Hurt
classify NeRF embeddings instead of rendering nine viewsResNet50 multi-view · NeRF classification inferencetotal time (ms): 97.56 -> 1.03Improved
interpolate nf2vec embeddings instead of NeRF weightsNeRF interpolation · ShapeNetRender interpolationvisual qualityImproved
use nf2vec embeddings instead of PointNet++ for retrievalpoint cloud retrieval · ModelNet40/ShapeNet10/ScanNet10mAP: average gap of 1.8 mAP vs PointNet++Mixed
use nf2vec embeddings instead of DGCNNpoint cloud classification · ModelNet40accuracy (%): 89.9 -> 87.0Hurt
use nf2vec instead of DWS on SDF classificationNF classification · Manifold40 (SDF)accuracy (%): 69.9 -> 86.3Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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