AI Paper Atlas

Research map / Point clouds / Segmentation and detection

Formula-Supervised Visual-Geometric Pre-training

Ryosuke Yamada, Kensho Hara, Hirokatsu Kataoka, Koshi Makihara, Nakamasa Inoue, Rio Yokota, Yutaka Satoh

ECCV 2024 · Point clouds › Segmentation and detection · approach: Multimodal and open-vocabulary 3D detection · 1 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
pre-train on both fractal images and point clouds (V+G) instead of visual only (ablation)FSVGP VG-FractalDB-1k · Image CLS average (7 datasets)accuracy: 89.3 -> 90.0Improved
pre-train on both modalities (V+G) instead of geometric only (ablation)FSVGP VG-FractalDB-1k · Image CLS average (7 datasets)accuracy: 67.2 -> 90.0Improved
add geometric modality to visual pre-training (ablation)FSVGP VG-FractalDB-1k · 3D CLS average (ModelNet40, ScanObjectNN)accuracy: 88.0 -> 88.3Improved
use formula-supervised consistency label instead of MAE self-supervision (ablation)VG-FractalDB-1k pre-training · ImageNet100accuracy: 80.3 -> 92.0Improved
use formula-supervised consistency label instead of MAE self-supervision (ablation)VG-FractalDB-1k pre-training · ModelNet40accuracy: 92.8 -> 92.9Improved
use fractal generation instead of Perlin noise (ablation)VG-FractalDB · ImageNet100accuracy: 90.7 -> 92.0Improved
use fractal generation instead of Perlin noise (ablation)VG-FractalDB · ModelNet40accuracy: 92.6 -> 92.9Improved
use fractal data instead of ShapeNet CAD projections at equal 50k size (ablation)VG-FractalDB · ImageNet100accuracy: 87.3 -> 87.9Improved
use fractal data instead of ShapeNet CAD projections at equal 50k size (ablation)VG-FractalDB · ModelNet40accuracy: 92.7 -> 92.8Improved
FSVGP with VG-FractalDB-1k versus VisualAtom-21k with 21x more dataFSVGP · Image CLS average (7 datasets)accuracy: 89.7 -> 90.0Improved
FSVGP versus real-data SSL pre-training (MAE, DINO) and supervised ImageNetFSVGP · Image CLS average (7 datasets)accuracy: 90.0 vs 92.5 (MAE)Hurt
FSVGP versus ShapeNet SSL methods for 3D classificationFSVGP · 3D CLS average (ModelNet40, ScanObjectNN)accuracy: 88.5 vs 89.5 (Point-BERT)Hurt

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