Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection
This work addresses the underutilization of 3D data in anomaly detection for computer vision applications, offering a high-performing solution that is not incremental but rather a novel combination of features.
The paper tackled the problem of effectively using 3D information for anomaly detection, finding that a simple 3D-only method combined with color features significantly outperforms previous state-of-the-art, achieving pixel-wise ROCAUC of 99.3% and PRO of 96.4% on the MVTec 3D-AD dataset.
Despite significant advances in image anomaly detection and segmentation, few methods use 3D information. We utilize a recently introduced 3D anomaly detection dataset to evaluate whether or not using 3D information is a lost opportunity. First, we present a surprising finding: standard color-only methods outperform all current methods that are explicitly designed to exploit 3D information. This is counter-intuitive as even a simple inspection of the dataset shows that color-only methods are insufficient for images containing geometric anomalies. This motivates the question: how can anomaly detection methods effectively use 3D information? We investigate a range of shape representations including hand-crafted and deep-learning-based; we demonstrate that rotation invariance plays the leading role in the performance. We uncover a simple 3D-only method that beats all recent approaches while not using deep learning, external pre-training datasets, or color information. As the 3D-only method cannot detect color and texture anomalies, we combine it with color-based features, significantly outperforming previous state-of-the-art. Our method, dubbed BTF (Back to the Feature) achieves pixel-wise ROCAUC: 99.3% and PRO: 96.4% on MVTec 3D-AD.