Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Fully Convolutional Cross-Scale-Flows for Image-based Defect Detection

About

In industrial manufacturing processes, errors frequently occur at unpredictable times and in unknown manifestations. We tackle the problem of automatic defect detection without requiring any image samples of defective parts. Recent works model the distribution of defect-free image data, using either strong statistical priors or overly simplified data representations. In contrast, our approach handles fine-grained representations incorporating the global and local image context while flexibly estimating the density. To this end, we propose a novel fully convolutional cross-scale normalizing flow (CS-Flow) that jointly processes multiple feature maps of different scales. Using normalizing flows to assign meaningful likelihoods to input samples allows for efficient defect detection on image-level. Moreover, due to the preserved spatial arrangement the latent space of the normalizing flow is interpretable which enables to localize defective regions in the image. Our work sets a new state-of-the-art in image-level defect detection on the benchmark datasets Magnetic Tile Defects and MVTec AD showing a 100% AUROC on 4 out of 15 classes.

Marco Rudolph, Tom Wehrbein, Bodo Rosenhahn, Bastian Wandt• 2021

Related benchmarks

TaskDatasetResultRank
Anomaly DetectionMVTec 3D-AD 1.0 (test)
Mean Score0.83
107
Anomaly DetectionMVTec AD
Overall AUROC98.7
83
Image-level Anomaly DetectionMVTec-AD (test)
Overall AUROC98.7
68
Anomaly DetectionMVTecAD (test)
Bottle Score100
55
Anomaly DetectionMVTec AD
Carpet AUROC100
40
Anomaly DetectionMTD
AUROC0.993
15
Anomaly DetectionMTD (test)
AUROC0.993
14
Image-level Anomaly DetectionMAD-Sim (test)
AUROC0.657
13
Anomaly DetectionMVTec AD 8 (test)
AUROC (Carpet)100
12
Image-level Anomaly DetectionMVTec SL (test)
Sensitivity (S)98.8
11
Showing 10 of 11 rows

Other info

Code

Follow for update