Skip to yearly menu bar Skip to main content


Poster

Odd-One-Out: Anomaly Detection by Comparing with Neighbors

Ankan Kumar Bhunia · Changjian Li · Hakan Bilen

ExHall D Poster #437
[ ] [ Paper PDF ]
[ Poster
Sat 14 Jun 3 p.m. PDT — 5 p.m. PDT

Abstract:

This paper introduces a novel anomaly detection (AD) problem aimed at identifying `odd-looking' objects within a scene by comparing them to other objects present. Unlike traditional AD benchmarks with fixed anomaly criteria, our task detects anomalies specific to each scene by inferring a reference group of regular objects. To address occlusions, we use multiple views of each scene as input, construct 3D object-centric models for each instance from 2D views, enhancing these models with geometrically consistent part-aware representations. Anomalous objects are then detected through cross-instance comparison. We also introduce two new benchmarks, ToysAD-8K and PartsAD-15K as testbeds for future research in this task. We provide a comprehensive analysis of our method quantitatively and qualitatively on these benchmarks. The datasets, source code, and models will be made publicly available upon publication.

Chat is not available.