A Streamlined System for Multimodal Industrial Anomaly Detection via 2D and 3D Feature Fusion

作者：Wenbing Zhu, Mingmin Chi, Bo Peng · 年份：2025 · DOI：10.1145/3746027.3761837 · 研究领域：Anomaly Detection Techniques and Applications、Fault Detection and Control Systems

We demonstrate an end-to-end system for real-time, multimodal industrial anomaly detection (IAD), built upon a custom hardware platform for synchronized 2D and 3D data acquisition. Our core contribution is a novel cross-modal residual mechanism that identifies defects by quantifying predictive errors between visual and geometric feature spaces. Instead of traditional concatenation, our dual-stream architecture mutually predicts features across modalities, leveraging the prediction residual's magnitude as a direct and robust anomaly indicator. The entire system achieves sub-second inference from acquisition to decision, enabled by efficient depth map analysis that circumvents the complexity of direct point cloud processing, offering a deployable solution for high-speed inspection.