Unsupervised Multi-View Visual Anomaly Detection via Progressive Homography-Guided Alignment

2025-11-24

AI Summary

Unsupervised multi-view visual anomaly detection faces a core challenge: distinguishing genuine defects from benign appearance variations induced by viewpoint changes. To address this, we propose ViewSense-AD—the first framework to incorporate geometric consistency modeling into this task. It introduces a homography-driven Multi-View Alignment Module (MVAM) for precise cross-view feature calibration; a View-aligned Latent Diffusion Model (VALDM) that progressively aligns multi-view representations in the latent space; and a lightweight Fusion Refinement Module (FRM) synergized with a memory bank, enabling anomaly discrimination via multi-level normal prototypes. Evaluated on RealIAD and MANTA benchmarks, ViewSense-AD achieves state-of-the-art performance, significantly reducing false positive rates while demonstrating strong robustness to large viewpoint variations and complex textures.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multi-instance/Multi-view Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

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
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Unsupervised visual anomaly detection from multi-view images presents a significant challenge: distinguishing genuine defects from benign appearance variations caused by viewpoint changes. Existing methods, often designed for single-view inputs, treat multiple views as a disconnected set of images, leading to inconsistent feature representations and a high false-positive rate. To address this, we introduce ViewSense-AD (VSAD), a novel framework that learns viewpoint-invariant representations by explicitly modeling geometric consistency across views. At its core is our Multi-View Alignment Module (MVAM), which leverages homography to project and align corresponding feature regions between neighboring views. We integrate MVAM into a View-Align Latent Diffusion Model (VALDM), enabling progressive and multi-stage alignment during the denoising process. This allows the model to build a coherent and holistic understanding of the object's surface from coarse to fine scales. Furthermore, a lightweight Fusion Refiner Module (FRM) enhances the global consistency of the aligned features, suppressing noise and improving discriminative power. Anomaly detection is performed by comparing multi-level features from the diffusion model against a learned memory bank of normal prototypes. Extensive experiments on the challenging RealIAD and MANTA datasets demonstrate that VSAD sets a new state-of-the-art, significantly outperforming existing methods in pixel, view, and sample-level visual anomaly proving its robustness to large viewpoint shifts and complex textures.

Problem

• Distinguishing genuine defects from benign appearance variations caused by viewpoint changes
• Addressing inconsistent feature representations in multi-view anomaly detection
• Reducing false-positive rates in unsupervised visual anomaly detection systems

Innovation

• Homography-guided multi-view feature alignment
• View-Align Latent Diffusion Model for progressive alignment
• Fusion Refiner Module for global consistency enhancement