Hierarchical Point-Patch Fusion with Adaptive Patch Codebook for 3D Shape Anomaly Detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kang, Xueyang, Li, Zizhao, Lan, Tian, Gong, Dong, Khoshelham, Kourosh, Nan, Liangliang
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915916637274112
author Kang, Xueyang
Li, Zizhao
Lan, Tian
Gong, Dong
Khoshelham, Kourosh
Nan, Liangliang
author_facet Kang, Xueyang
Li, Zizhao
Lan, Tian
Gong, Dong
Khoshelham, Kourosh
Nan, Liangliang
contents 3D shape anomaly detection is a crucial task for industrial inspection and geometric analysis. Existing deep learning approaches typically learn representations of normal shapes and identify anomalies via out-of-distribution feature detection or decoder-based reconstruction. They often fail to generalize across diverse anomaly types and scales, such as global geometric errors (e.g., planar shifts, angle misalignments), and are sensitive to noisy or incomplete local points during training. To address these limitations, we propose a hierarchical point-patch anomaly scoring network that jointly models regional part features and local point features for robust anomaly reasoning. An adaptive patchification module integrates self-supervised decomposition to capture complex structural deviations. Beyond evaluations on public benchmarks (Anomaly-ShapeNet and Real3D-AD), we release an industrial test set with real CAD models exhibiting planar, angular, and structural defects. Experiments on public and industrial datasets show superior AUC-ROC and AUC-PR performance, including over 40% point-level improvement on the new industrial anomaly type and average object-level gains of 7% on Real3D-AD and 4% on Anomaly-ShapeNet, demonstrating strong robustness and generalization.
format Preprint
id arxiv_https___arxiv_org_abs_2604_03972
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Hierarchical Point-Patch Fusion with Adaptive Patch Codebook for 3D Shape Anomaly Detection
Kang, Xueyang
Li, Zizhao
Lan, Tian
Gong, Dong
Khoshelham, Kourosh
Nan, Liangliang
Computer Vision and Pattern Recognition
68T07 (Primary), 68T45, 68U05, 68T09 (Secondary)
I.2.10; I.4.8; I.5.1; I.3.5
3D shape anomaly detection is a crucial task for industrial inspection and geometric analysis. Existing deep learning approaches typically learn representations of normal shapes and identify anomalies via out-of-distribution feature detection or decoder-based reconstruction. They often fail to generalize across diverse anomaly types and scales, such as global geometric errors (e.g., planar shifts, angle misalignments), and are sensitive to noisy or incomplete local points during training. To address these limitations, we propose a hierarchical point-patch anomaly scoring network that jointly models regional part features and local point features for robust anomaly reasoning. An adaptive patchification module integrates self-supervised decomposition to capture complex structural deviations. Beyond evaluations on public benchmarks (Anomaly-ShapeNet and Real3D-AD), we release an industrial test set with real CAD models exhibiting planar, angular, and structural defects. Experiments on public and industrial datasets show superior AUC-ROC and AUC-PR performance, including over 40% point-level improvement on the new industrial anomaly type and average object-level gains of 7% on Real3D-AD and 4% on Anomaly-ShapeNet, demonstrating strong robustness and generalization.
title Hierarchical Point-Patch Fusion with Adaptive Patch Codebook for 3D Shape Anomaly Detection
topic Computer Vision and Pattern Recognition
68T07 (Primary), 68T45, 68U05, 68T09 (Secondary)
I.2.10; I.4.8; I.5.1; I.3.5
url https://arxiv.org/abs/2604.03972