CASL: Curvature-Augmented Self-supervised Learning for 3D Anomaly Detection

2025-11-16

AI Summary

Existing 3D anomaly detection methods suffer from poor generalization, while generic self-supervised models underperform in anomaly detection. To address this, we propose a curvature-enhanced self-supervised point cloud learning framework. Our method innovatively incorporates multi-scale curvature as prompt signals into coordinate reconstruction, guiding the decoder to focus on geometrically anomalous regions without task-specific architectural modifications. Leveraging a U-Net-based autoencoder, we perform curvature-prompted self-supervised pretraining; downstream tasks—including anomaly detection and point cloud classification—are then unified via lightweight task heads. On mainstream 3D anomaly detection benchmarks, our approach significantly outperforms both dedicated detectors and classical self-supervised methods. Moreover, it maintains strong generalization on classification benchmarks such as ModelNet40. This framework achieves a synergistic improvement in detection accuracy and model versatility, bridging the gap between specialized detection performance and broad applicability.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: 3D Computer Vision
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Deep learning-based 3D anomaly detection methods have demonstrated significant potential in industrial manufacturing. However, many approaches are specifically designed for anomaly detection tasks, which limits their generalizability to other 3D understanding tasks. In contrast, self-supervised point cloud models aim for general-purpose representation learning, yet our investigation reveals that these classical models are suboptimal at anomaly detection under the unified fine-tuning paradigm. This motivates us to develop a more generalizable 3D model that can effectively detect anomalies without relying on task-specific designs. Interestingly, we find that using only the curvature of each point as its anomaly score already outperforms several classical self-supervised and dedicated anomaly detection models, highlighting the critical role of curvature in 3D anomaly detection. In this paper, we propose a Curvature-Augmented Self-supervised Learning (CASL) framework based on a reconstruction paradigm. Built upon the classical U-Net architecture, our approach introduces multi-scale curvature prompts to guide the decoder in predicting the spatial coordinates of each point. Without relying on any dedicated anomaly detection mechanisms, it achieves leading detection performance through straightforward anomaly classification fine-tuning. Moreover, the learned representations generalize well to standard 3D understanding tasks such as point cloud classification. The code is available at https://github.com/zyh16143998882/CASL.

Problem

• Enhancing 3D anomaly detection using curvature-augmented self-supervised learning
• Overcoming limited generalizability in specialized 3D anomaly detection methods
• Improving classical self-supervised models for unified 3D understanding tasks

Innovation

• Curvature prompts guide U-Net reconstruction
• Multi-scale curvature augments self-supervised learning
• Anomaly detection via classification fine-tuning