# DCRDF-Net: A Dual-Channel Reverse-Distillation Fusion Network for 3D Industrial Anomaly Detection

## TL;DR

This paper introduces DCRDF-Net, a new method for detecting defects in industrial products using RGB and depth data, even when few defective samples are available.

## Contribution

DCRDF-Net introduces a novel fusion network with reverse-distillation and pseudo-anomaly generation for unsupervised 3D industrial anomaly detection.

## Key findings

- DCRDF-Net achieves 97.1% image-level I-AUROC on the MVTec 3D-AD dataset.
- The method attains 98.8% pixel-level PRO, outperforming existing multimodal approaches.
- The framework effectively handles cross-modal inconsistencies and modality-specific features.

## Abstract

Industrial surface defect detection is essential for ensuring product quality, but real-world production lines often provide only a limited number of defective samples, making supervised training difficult. Multimodal anomaly detection with aligned RGB and depth data is a promising solution, yet existing fusion schemes tend to overlook modality-specific characteristics and cross-modal inconsistencies, so that defects visible in only one modality may be suppressed or diluted. In this work, we propose DCRDF-Net, a dual-channel reverse-distillation fusion network for unsupervised RGB–depth industrial anomaly detection. The framework learns modality-specific normal manifolds from nominal RGB and depth data and detects defects as deviations from these learned manifolds. It consists of three collaborative components: a Perlin-guided pseudo-anomaly generator that injects appearance–geometry-consistent perturbations into both modalities to enrich training signals; a dual-channel reverse-distillation architecture with guided feature refinement that denoises teacher features and constrains RGB and depth students towards clean, defect-free representations; and a cross-modal squeeze–excitation gated fusion module that adaptively combines RGB and depth anomaly evidence based on their reliability and agreement.Extensive experiments on the MVTec 3D-AD dataset show that DCRDF-Net achieves 97.1% image-level I-AUROC and 98.8% pixel-level PRO, surpassing current state-of-the-art multimodal methods on this benchmark.