论文简读Multi-scale feature reconstruction network for industrial anomaly detection

Multi-scale feature reconstruction network for industrial anomaly detection

E. Iqbal, S.U. Khan, S. Javed et al., Multi-scale feature reconstruction network for industrial anomaly detection,
Knowledge-Based Systems (2024), doi: https://doi.org/10.1016/j.knosys.2024.112650.

思路在于训练阶段输入正常样本，图中的Masked Vision encoder 通过掩码训练 重建出 ResNet的多阶段输出，然而在测试阶段Masked Vision encoder 输出正常的重建特征，而ResNet的输出会有缺陷存在。通过这样方式进行定位异常