Towards Scalable IoT Deployment for Visual Anomaly Detection via Efficient Compression

TL;DR

This paper explores scalable visual anomaly detection in IoT environments by employing efficient data compression techniques, achieving significant latency reduction with minimal accuracy loss in industrial settings.

Contribution

It introduces a comprehensive evaluation of compression strategies tailored for IoT-based VAD, demonstrating their effectiveness in reducing latency while maintaining detection performance.

Findings

Up to 80% reduction in inference time.

Significant compression with minimal accuracy loss.

Effective tradeoff between latency and detection accuracy.

Abstract

Visual Anomaly Detection (VAD) is a key task in industrial settings, where minimizing operational costs is essential. Deploying deep learning models within Internet of Things (IoT) environments introduces specific challenges due to limited computational power and bandwidth of edge devices. This study investigates how to perform VAD effectively under such constraints by leveraging compact, efficient processing strategies. We evaluate several data compression techniques, examining the tradeoff between system latency and detection accuracy. Experiments on the MVTec AD benchmark demonstrate that significant compression can be achieved with minimal loss in anomaly detection performance compared to uncompressed data. Current results show up to 80% reduction in end-to-end inference time, including edge processing, transmission, and server computation.