mlsanityData to 14 Oct 2026

Understanding and Overcoming Cross-modal Fusion Bias in Multimodal Anomaly Detection From A Fisher Information Perspective

arXiv 2608.00986 ↗First posted 2 Aug 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Current advancements in Multimodal Anomaly Detection (MAD) are largely driven by enhancing multimodal fusion, particularly through the integration of RGB and Depth data for richer anomaly representation. However, less attention was devoted to analyzing the role of cross-modal fusion bias, a well-known challenge in multimodal learning, in MAD. This gap motivates a key question: can we overcome this bias to break the performance bottleneck of current work? In this paper, we first analyze the impact of cross-modal fusion bias in MAD via the Fisher Information Matrix. Then, grounded in these findings, we propose UCFB, a simple yet effective plug-and-play framework designed to mitigate cross-modal fusion bias in MAD. It achieves this by jointly employing Fisher-information-guided dynamic calibration to adjust modality-specific regularization weights and canonical similarity analysis to improve inter-modal interactions. Extensive experiments on the MVTec 3D-AD and Eyecandies datasets demonstrate that UCFB achieves consistent improvements in single-class, multi-class, and few-shot settings.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 39 of its 69 references that are in the map. First posted on arXiv 2 Aug 2026.

Builds on

39 of 69 references.

References · 39

  1. 2020PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localizationcited by 594
  2. 2021CutPaste: Self-Supervised Learning for Anomaly Detection and Localizationcited by 531
  3. 2021DRAEM -- A discriminatively trained reconstruction embedding for surface anomaly detectioncited by 514
  4. 2022Anomaly Detection via Reverse Distillation from One-Class Embeddingcited by 421
  5. 2022A Unified Model for Multi-class Anomaly Detectioncited by 268
  6. 2022Balanced Multimodal Learning via On-the-fly Gradient Modulationcited by 264
  7. 2023Multimodal Industrial Anomaly Detection via Hybrid Fusioncited by 139
  8. 2024AdaCLIP: Adapting CLIP with Hybrid Learnable Prompts for Zero-Shot Anomaly Detectioncited by 127
  9. 2021Fully Convolutional Cross-Scale-Flows for Image-based Defect Detectioncited by 117
  10. 2024RealNet: A Feature Selection Network with Realistic Synthetic Anomaly for Anomaly Detectioncited by 116
  11. 2023AnomalyDiffusion: Few-Shot Anomaly Image Generation with Diffusion Modelcited by 111
  12. 2022Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detectioncited by 93
27 more
  1. 2022DSR -- A dual subspace re-projection network for surface anomaly detectioncited by 90
  2. 2023Towards Scalable 3D Anomaly Detection and Localization: A Benchmark via 3D Anomaly Synthesis and A Self-Supervised Learning Networkcited by 67
  3. 2024SoftPatch: Unsupervised Anomaly Detection with Noisy Datacited by 55
  4. 2025Frequency-Aligned Knowledge Distillation for Lightweight Spatiotemporal Forecastingcited by 55
  5. 2023Multimodal Industrial Anomaly Detection by Crossmodal Feature Mappingcited by 42
  6. 2026ReTrack: Evidence-Driven Dual-Stream Directional Anchor Calibration Network for Composed Video Retrievalcited by 41
  7. 2023Cheating Depth: Enhancing 3D Surface Anomaly Detection via Depth Simulationcited by 33
  8. 2024On-the-fly Modulation for Balanced Multimodal Learningcited by 32
  9. 2024GeneralAD: Anomaly Detection Across Domains by Attending to Distorted Featurescited by 29
  10. 2024SeaS: Few-shot Industrial Anomaly Image Generation with Separation and Sharing Fine-tuningcited by 27
  11. 2024Multi-Sensor Object Anomaly Detection: Unifying Appearance, Geometry, and Internal Propertiescited by 26
  12. 2025MC3D-AD: A Unified Geometry-aware Reconstruction Model for Multi-category 3D Anomaly Detectioncited by 23
  13. 2023Target before Shooting: Accurate Anomaly Detection and Localization under One Millisecond via Cascade Patch Retrievalcited by 22
  14. 2025One-for-More: Continual Diffusion Model for Anomaly Detectioncited by 19
  15. 2025Adaptive Unimodal Regulation for Balanced Multimodal Information Acquisitioncited by 19
  16. 2025Fishers for Free? Approximating the Fisher Information Matrix by Recycling the Squared Gradient Accumulatorcited by 14
  17. 2025MultiADS: Defect-aware Supervision for Multi-type Anomaly Detection and Segmentation in Zero-Shot Learningcited by 12
  18. 2026COMBINER: Composed Image Retrieval Guided by Attribute-based Neighbor Relationscited by 11
  19. 2025Bridging 3D Anomaly Localization and Repair via High-Quality Continuous Geometric Representationcited by 9
  20. 2026TP-Seg: Task-Prototype Framework for Unified Medical Lesion Segmentationcited by 7
  21. 2025Commonality in Few: Few-Shot Multimodal Anomaly Detection via Hypergraph-Enhanced Memorycited by 6
  22. 2025Modality-Balanced Collaborative Distillation for Multi-Modal Domain Generalizationcited by 6
  23. 2026DifferSeg: Towards Diverse Multimodal Binary Segmentation via Differential Perception and Frequency Guidancecited by 6
  24. 2025CostFilter-AD: Enhancing Anomaly Detection through Matching Cost Filteringcited by 5
  25. 2026Seeing the End at Step Zero: Accelerating Diffusion MLLMs via MLP Sparsity-Aware Truncationcited by 5
  26. 2026HVPNet: A Bio-Inspired Network for General Salient and Camouflaged Object Detectioncited by 4
  27. 2024Robust Modality-incomplete Anomaly Detection: A Modality-instructive Framework with Benchmarkcited by 3

Built on by

Not cited yet.