Hierarchical Point-Patch Fusion with Adaptive Patch Codebook for 3D Shape Anomaly Detection
Xueyang Kang, Zizhao Li, Tian Lan, Dong Gong, Kourosh Khoshelham, Liangliang Nan, [0.5em] The University of Melbourne, Parkville, Victoria 3010, Australia · 2026
Read Hierarchical Point-Patch Fusion with Adaptive Patch Codebook for 3D Shape Anomaly Detection by Xueyang Kang, Zizhao Li, Tian Lan, Dong Gong, Kourosh Khoshelham, Liangliang Nan, [0.5em] The University of Melbourne, Parkville, Victoria 3010, Australia (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-03972.txt.
Description
Xueyang Kang and colleagues built a hierarchical point-patch network for 3D shape anomaly detection in industrial inspection. Earlier methods learn normal shapes and flag outliers, but struggle across anomaly types and scales and falter on noisy local points. Their network handles regional part features together with local point features, with an adaptive patch module decomposing shapes by self-supervision. Tests on public sets plus a new industrial CAD set showed strong gains, including over 40 percent point-level improvement on the new anomaly type.
Catalog record
| CREATOR | Xueyang Kang, Zizhao Li, Tian Lan, Dong Gong, Kourosh Khoshelham, Liangliang Nan, [0.5em] The University of Melbourne, Parkville, Victoria 3010, Australia |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39488 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.03972