MVFM-3DAD: Multi-view Flow Matching for 3D Anomaly Detection via Density Proxy Estimation

2026-08-12

AI Summary

Existing 3D anomaly detection methods struggle to accurately model the distribution of normal data due to insufficient memory bank coverage or identity shortcut issues in reconstruction. This work proposes a reconstruction-free and explicit Jacobian computation-free flow matching framework, reframing 3D anomaly detection as a density proxy estimation task. Specifically, a Bidirectional Geometric Projector (BGP) generates multi-view representations, and a Flow-guided Density Proxy Estimator (FDPE) evaluates the normality of features from each view, which are subsequently mapped back into 3D space for anomaly discrimination. The proposed method achieves significant performance gains over current state-of-the-art approaches on both Real3D-AD and MVTec3D-AD benchmarks, demonstrating its effectiveness and superiority.

Abstract

In 3D anomaly detection (3DAD), most existing methods rely on Memory bank retrieval or reconstruction. However, memory-based methods are constrained by the coverage of stored normal features, while reconstruction-based methods may learn identity shortcuts that also reconstruct anomalous inputs well. These limitations motivate a density-oriented approach that evaluates whether a test sample follows the learned normal distribution. To this end, we propose MVFM-3DAD, a flow-based framework that reframes 3DAD as density proxy estimation over the normal data distribution. MVFM-3DAD introduces a Bidirectional Geometric Projector (BGP), whose forward process converts irregular point clouds into structured multi-view representations. The Flow-guided Density Proxy Estimator (FDPE) estimates a reference density for each view feature, after which the backward process of BGP maps these multi-view density estimates to their corresponding 3D points. Building on it, anomalous features can be identified by their terminal normality. Unlike conventional flow-based likelihood estimation, our formulation requires neither input reconstruction nor explicit Jacobian evaluation, yielding a simple and efficient anomaly-scoring mechanism. Extensive experiments show that MVFM-3DAD outperforms the strongest competing methods on Real3D-AD and MVTec3D-AD. Code is available at https://github.com/lil-wayne-0319/MV3D-AD

Problem

• 3D anomaly detection
• density estimation
• point cloud
• normality assessment
• anomaly scoring

Innovation

• flow-based anomaly detection
• density proxy estimation
• multi-view representation
• 3D point cloud
• normality scoring