用金字塔扫描策略提升小异常定位精度
Pyramid-based Mamba Multi-class Unsupervised Anomaly Detection
- 基于状态空间模型的金字塔扫描策略，多尺度提取特征
- 在MVTec上实现+1%的AP和+1%的AU-PRO提升
- 适合工业场景中细粒度异常检测与定位任务
卷积神经网络（CNN）和基于Transformer的方法虽提升了异常检测与定位性能，但在精确定位小异常方面仍面临挑战。CNN难以捕捉长距离依赖，而Transformer则存在计算开销大的问题。本文提出一种基于状态空间模型（SSM）的金字塔扫描策略（PSS），用于多类异常检测与定位。该方法通过将PSS与预训练编码器结合，实现多尺度特征提取，并引入特征级合成异常生成器。在MVTec基准测试中，多类异常定位的AP提升1%，AU-PRO提高1%，验证了其在多样化工业场景中对小异常精确定位的优势。代码已开源：https://github.com/iqbalmlpuniud/Pyramid Mamba。
原文摘要 · Abstract (English)
Recent advances in convolutional neural networks (CNNs) and transformer-based methods have improved anomaly detection and localization, but challenges persist in precisely localizing small anomalies. While CNNs face limitations in capturing long-range dependencies, transformer architectures often suffer from substantial computational overheads. We introduce a state space model (SSM)-based Pyramidal Scanning Strategy (PSS) for multi-class anomaly detection and localization--a novel approach designed to address the challenge of small anomaly localization. Our method captures fine-grained details at multiple scales by integrating the PSS with a pre-trained encoder for multi-scale feature extraction and a feature-level synthetic anomaly generator. An improvement of $+1\%$ AP for multi-class anomaly localization and a +$1\%$ increase in AU-PRO on MVTec benchmark demonstrate our method's superiority in precise anomaly localization across diverse industrial scenarios. The code is available at https://github.com/iqbalmlpuniud/Pyramid Mamba.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。