Saved in:
Bibliographic Details
Main Authors: Bhunia, Ankan, Li, Changjian, Bilen, Hakan
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2406.20099
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909547630690304
author Bhunia, Ankan
Li, Changjian
Bilen, Hakan
author_facet Bhunia, Ankan
Li, Changjian
Bilen, Hakan
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2406_20099
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Odd-One-Out: Anomaly Detection by Comparing with Neighbors
Bhunia, Ankan
Li, Changjian
Bilen, Hakan
Computer Vision and Pattern Recognition
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.
title Odd-One-Out: Anomaly Detection by Comparing with Neighbors
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2406.20099