Research map / 3D vision / 3D generation
NOVUM: Neural Object Volumes for Robust Object Classification
Artur Jesslen, Guofeng Zhang, Angtian Wang, Wufei Ma, Alan Yuille, Adam Kortylewski
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 changed | Where | Measured | Effect |
|---|---|---|---|
| replace 3D Gaussian cuboid volume with a single feature vector per class (ablation) | NOVUM · P3D+ L0 and occluded-P3D+ L3 | classification accuracy (%): 99.5 -> 93.2 (L0); 59.2 -> 44.0 (L3) | Hurt |
| add background features during training (ablation) | NOVUM · occluded-P3D+ L3 | classification accuracy (%): 53.0 -> 59.2 | Improved |
| add background features during training (ablation) | NOVUM · P3D+ L0 (IID) | classification accuracy (%): 99.3 -> 99.5 | No clear effect |
| use cuboid instead of sphere Gaussian layout (ablation) | NOVUM · occluded-P3D+ L3 | classification accuracy (%): 59.0 -> 59.2 | No clear effect |
| use neural object volumes instead of feed-forward classifiers | NOVUM · OOD-CV | classification accuracy (%): 64.2 -> 85.2 (vs Swin-T) | Improved |
| use neural object volumes instead of feed-forward classifiers | NOVUM · occluded-P3D+ mean | classification accuracy (%): 75.8 -> 81.6 (vs Convnext) | Improved |
| use neural object volumes instead of feed-forward classifiers on unoccluded data | NOVUM · P3D+ L0 | classification accuracy (%): 99.4 -> 99.5 | No clear effect |
| use NOVUM volume rendering instead of NeMo mesh render-and-compare for pose | NOVUM · P3D+ 3D pose pi/6 | accuracy (%): 86.1 -> 88.2 | Improved |
| use NOVUM instead of NeMo for fine pose threshold | NOVUM · P3D+ 3D pose pi/18 | accuracy (%): 61.0 -> 59.1 | Mixed |
| use class-agnostic backbone and neural volumes instead of per-class NeMo models | NOVUM · pose inference efficiency | parameters / optimization steps: 996M -> 83M params; 300 -> 30 steps | Improved |
| increase parameter count among baselines | classification baselines · OOD robustness | accuracy: Swin-T 28M, ViT-b-16 86M, NOVUM 83M | No clear effect |
Later papers that compared with this method
2 later accepted papers report NOVUM as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Interpretable 3D Neural Object Volumes for Robust Conceptual Reasoning (ICLR 2026 · better on 2 of 4 shared results)
- Scaling 3D Compositional Models for Robust Classification and Pose Estimation (ICCV 2025 · better on 7 of 7 shared results)
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.
- NeMo — NeMo: Learning 3D Neural Motion Fields From Multiple Video Instances of the Same Action (CVPR 2023) · better on 3 of 4 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
- Scaling 3D Compositional Models for Robust Classification and Pose Estimation (ICCV 2025 · similarity 0.89)
- Compositional Convolutional Neural Networks: A Robust and Interpretable Model for Object Recognition Under Occlusion (IJCV 2020 · similarity 0.88)
- Contemplating Real-World Object Classification (ICLR 2021 · similarity 0.87)
- Survey and Evaluation of Neural 3D Shape Classification Approaches (TPAMI 2021 · similarity 0.87)
- Volumetric and Multi-View CNNs for Object Classification on 3D Data (CVPR 2016 · similarity 0.87)
- Compositional Convolutional Neural Networks: A Deep Architecture With Innate Robustness to Partial Occlusion (CVPR 2020 · similarity 0.87)
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.
