EdgeZSAD: Practical Zero-Shot Anomaly Detection on Edge Devices
2026-06-15 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors address the challenge of detecting anomalies in industrial images using models small enough to run on edge devices with limited memory. They propose EdgeZSAD, which uses a compact TinyViT backbone and a special readout method to keep the model lightweight and efficient. Their model is trained on known data but tested on unseen targets across six benchmarks, showing strong performance while running smoothly on real embedded hardware. The results suggest their approach maintains accuracy without needing large, resource-heavy models.
Zero-shot anomaly detectionIndustrial inspectionEdge deploymentTinyViTAUROCJetson Orin NanoTensorRT FP16Global-local readoutBenchmarkingModel compression
Authors
Taewan Cho, Andrew Jaeyong Choi
Abstract
Industrial inspection needs zero-shot anomaly detection (ZSAD) that remains useful under edge deployment constraints. Recent methods often rely on ViT-L foundation backbones (~300M parameters), which exceed the memory and operator budget of typical embedded hardware. We study this regime through EdgeZSAD, a compact reference system built around a TinyViT-21M-512 backbone, an asymmetric global-local readout (EdgeGLR), and a reproducible source-side training recipe (Real-IAD-DR). We train a single checkpoint in a source-trained, target-unseen protocol and evaluate it across six industrial benchmarks. Across three independent runs, the resulting model reaches an average image AUROC of 91.6 on MVTec-AD and 88.2 on VisA, while remaining directly deployable on Jetson Orin Nano Super (TensorRT FP16) and RB5 Gen2 (QNN GPU FP16). Across the six device-rescored benchmarks, image-AUROC drift stays below 0.2 points, indicating that the exported graph preserves host-side ranking behavior in the evaluated deployment setting.