arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2503.23451cs.CV
收藏

超越学术基准:视觉工业异常检测的批判性分析与最佳实践

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

Valeo(法雷奥);Intel(英特尔)

Aimira Baitieva, Yacine Bouaouni, Alexandre Briot, Dick Ameln, Souhaiel Khalfaoui, Samet Akcay

AI总结:

本文批判性分析视觉工业异常检测,强调真实数据基准与实用指标,提出三项贡献以弥合学术与工业差距。

摘要:

异常检测(AD)对于制造业中的自动化视觉检测至关重要。这一计算机视觉领域正在快速发展,对实际应用的关注日益增多。与此同时,流行的数据集通常在受控的实验室环境中生成,缺陷为人工制造,无法捕捉真实生产条件的多样性。新方法在生产环境中往往失效,表现出显著的性能下降或需要不切实际的计算资源。学术成果与工业可行性之间的脱节威胁着视觉异常检测研究的正确方向。本文做出三项关键贡献:(1)我们证明了真实世界数据集的重要性,并使用实际生产数据建立了基准;(2)我们利用对实际应用有价值的指标,对现有SOTA方法在多样化任务上进行了公平比较;(3)我们通过讨论重要挑战和弥合学术界与工业界差距的新视角,对该领域的最新进展进行了全面分析。代码公开于https://github.com/abc-125/viad-benchmark

英文摘要:

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

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑