Divide and Conquer: High-Resolution Industrial Anomaly Detection via Memory Efficient Tiled Ensemble

TL;DR

This paper introduces a tiled ensemble approach for high-resolution industrial anomaly detection that significantly reduces memory usage by training models on image tiles, enabling effective detection without extensive hardware requirements.

Contribution

The proposed method allows high-resolution anomaly detection with minimal memory overhead by dividing images into overlapping tiles and training dedicated models for each, compatible with existing architectures.

Findings

Improves anomaly detection accuracy across multiple architectures.

Reduces GPU memory consumption to that of processing a single tile.

Achieves better results on MVTec and VisA datasets.

Abstract

Industrial anomaly detection is an important task within computer vision with a wide range of practical use cases. The small size of anomalous regions in many real-world datasets necessitates processing the images at a high resolution. This frequently poses significant challenges concerning memory consumption during the model training and inference stages, leaving some existing methods impractical for widespread adoption. To overcome this challenge, we present the tiled ensemble approach, which reduces memory consumption by dividing the input images into a grid of tiles and training a dedicated model for each tile location. The tiled ensemble is compatible with any existing anomaly detection model without the need for any modification of the underlying architecture. By introducing overlapping tiles, we utilize the benefits of traditional stacking ensembles, leading to further improvements in anomaly detection capabilities beyond high resolution alone. We perform a comprehensive analysis using diverse underlying architectures, including Padim, PatchCore, FastFlow, and Reverse Distillation, on two standard anomaly detection datasets: MVTec and VisA. Our method demonstrates a notable improvement across setups while remaining within GPU memory constraints, consuming only as much GPU memory as a single model needs to process a single tile.