VT-3DAD fuses visual deviation from few-shot references and semantic deviation from textual normal space to achieve SOTA cross-category 3D anomaly detection on ShapeNetPart.
Towards zero-shot point cloud anomaly detection: A multi-view projection framework,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
VT-3DAD: Cross-Category 3D Anomaly Detection via Visual-Text Normal Space Alignment
VT-3DAD fuses visual deviation from few-shot references and semantic deviation from textual normal space to achieve SOTA cross-category 3D anomaly detection on ShapeNetPart.