AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

Hyperbolic Defect Feature Synthesis for Few-Shot Defect Classification

Huimin Li, Boxuan Hu, Yulin Zhang, Xiuzhuang Zhou, Junlin Hu

CVPR 2026 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Fine-grained and cross-domain few-shot baselines

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 AlphaCLIP mask weighting instead of equal mask weight (ablation)AlphaCLIP ViT-L/14 · MTD 5-shotmean accuracy: 51.9 -> 69.5Improved
use ViT backbone instead of ResNet (ablation)VanillaClip · MTD 5-shotmean accuracy: 24.3 (ResNet50) -> 52.7 (ViT-B/16)Improved
use AlphaCLIP instead of vanilla CLIP (ablation)ViT-L/14 · MTD 5-shotmean accuracy: 52.5 -> 69.5Improved
set hyperbolic curvature c=0.1 instead of Euclidean c=0 (ablation)HypDFS · MTD 5-shotmean accuracy: 78.1 -> 79.1Improved
make curvature learnable (ablation)HypDFS · MTD 5-shotmean accuracy: 79.2 -> 78.9No clear effect
vary fixed curvature from 0.01 to 1.0 (ablation)HypDFS · MTD 5-shotmean accuracy: 78.9 to 79.2No clear effect
add residual adapter trained on real features (ablation)HypDFS · MTD 5-shotmean accuracy: 69.5 -> 78.1Improved
add synthetic hyperbolic defect features to adapter training (ablation)HypDFS · MTD 5-shotmean accuracy: 78.1 -> 86.4Improved
add hierarchical defect contrastive loss in hyperbolic space (ablation)HypDFS · MTD 5-shotmean accuracy: 79.2 -> 86.4Improved
use hyperbolic distance without synthetic features (ablation)HypDFS · MTD 5-shotmean accuracy: 78.1 -> 79.2Improved
replace Euclidean synthesis with hyperbolic synthesisZip-A-F · MTD 3-shotmean accuracy: 67.6 -> 78.6Improved

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