Beyond Academic Benchmarks: Critical Analysis and Best Practices for Visual Industrial Anomaly Detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Baitieva, Aimira, Bouaouni, Yacine, Briot, Alexandre, Ameln, Dick, Khalfaoui, Souhaiel, Akcay, Samet
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913766596149248
author Baitieva, Aimira
Bouaouni, Yacine
Briot, Alexandre
Ameln, Dick
Khalfaoui, Souhaiel
Akcay, Samet
author_facet Baitieva, Aimira
Bouaouni, Yacine
Briot, Alexandre
Ameln, Dick
Khalfaoui, Souhaiel
Akcay, Samet
contents Anomaly detection (AD) is essential for automating visual inspection in manufacturing. This field of computer vision is rapidly evolving, with increasing attention towards real-world applications. Meanwhile, popular datasets are typically produced in controlled lab environments with artificially created defects, unable to capture the diversity of real production conditions. New methods often fail in production settings, showing significant performance degradation or requiring impractical computational resources. This disconnect between academic results and industrial viability threatens to misdirect visual anomaly detection research. This paper makes three key contributions: (1) we demonstrate the importance of real-world datasets and establish benchmarks using actual production data, (2) we provide a fair comparison of existing SOTA methods across diverse tasks by utilizing metrics that are valuable for practical applications, and (3) we present a comprehensive analysis of recent advancements in this field by discussing important challenges and new perspectives for bridging the academia-industry gap. The code is publicly available at https://github.com/abc-125/viad-benchmark
format Preprint
id arxiv_https___arxiv_org_abs_2503_23451
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Beyond Academic Benchmarks: Critical Analysis and Best Practices for Visual Industrial Anomaly Detection
Baitieva, Aimira
Bouaouni, Yacine
Briot, Alexandre
Ameln, Dick
Khalfaoui, Souhaiel
Akcay, Samet
Computer Vision and Pattern Recognition
Anomaly detection (AD) is essential for automating visual inspection in manufacturing. This field of computer vision is rapidly evolving, with increasing attention towards real-world applications. Meanwhile, popular datasets are typically produced in controlled lab environments with artificially created defects, unable to capture the diversity of real production conditions. New methods often fail in production settings, showing significant performance degradation or requiring impractical computational resources. This disconnect between academic results and industrial viability threatens to misdirect visual anomaly detection research. This paper makes three key contributions: (1) we demonstrate the importance of real-world datasets and establish benchmarks using actual production data, (2) we provide a fair comparison of existing SOTA methods across diverse tasks by utilizing metrics that are valuable for practical applications, and (3) we present a comprehensive analysis of recent advancements in this field by discussing important challenges and new perspectives for bridging the academia-industry gap. The code is publicly available at https://github.com/abc-125/viad-benchmark
title Beyond Academic Benchmarks: Critical Analysis and Best Practices for Visual Industrial Anomaly Detection
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.23451