MoECLIP: Patch-Specialized Experts for Zero-shot Anomaly Detection

arXiv 2603.03101 ↗First posted 3 Mar 2026cs.CV cs.AIcited by 0

On the map: Mixture-of-experts → Vision mixture-of-experts

Abstract (arXiv)

The CLIP model's outstanding generalization has driven recent success in Zero-Shot Anomaly Detection (ZSAD) for detecting anomalies in unseen categories. The core challenge in ZSAD is to specialize the model for anomaly detection tasks while preserving CLIP's powerful generalization capability. Existing approaches attempting to solve this challenge share the fundamental limitation of a patch-agnostic design that processes all patches monolithically without regard for their unique characteristics. To address this limitation, we propose MoECLIP, a Mixture-of-Experts (MoE) architecture for the ZSAD task, which achieves patch-level adaptation by dynamically routing each image patch to a specialized Low-Rank Adaptation (LoRA) expert based on its unique characteristics. Furthermore, to prevent functional redundancy among the LoRA experts, we introduce (1) Frozen Orthogonal Feature Separation (FOFS), which orthogonally separates the input feature space to force experts to focus on distinct information, and (2) a simplex equiangular tight frame (ETF) loss to regulate the expert outputs to form maximally equiangular representations. Comprehensive experimental results across 14 benchmark datasets spanning industrial and medical domains demonstrate that MoECLIP outperforms existing state-of-the-art methods. The code is available at https://github.com/CoCoRessa/MoECLIP.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Vision mixture-of-experts, one of 12 directions in Mixture-of-experts. Builds on 42 of its 58 references that are in the map. First posted on arXiv 3 Mar 2026.

Builds on

42 of 58 references.

Closest ancestors · 1

Same direction, newest first

2025AnomalyMoE: Towards a Language-free Generalist Model for Unified Visual Anomaly Detection8 Aug

Other references · 41

2016Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localizationcited by 6,690
2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,827
2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,413
2021Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsitycited by 3,014
2016V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentationcited by 2,284
2024Mixtral of Expertscited by 1,785
2020GShard: Scaling Giant Models with Conditional Computation and Automatic Shardingcited by 1,601
2021GLaM: Efficient Scaling of Language Models with Mixture-of-Expertscited by 890
2021Towards Total Recall in Industrial Anomaly Detectioncited by 772
2020Prevalence of Neural Collapse during the terminal phase of deep learning trainingcited by 713
2024Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Surveycited by 641
2020PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localizationcited by 591

29 more

2019Kvasir-SEG: A Segmented Polyp Datasetcited by 588
2022Anomaly Detection via Reverse Distillation from One-Class Embeddingcited by 419
2022SPot-the-Difference Self-Supervised Pre-training for Anomaly Detection and Segmentationcited by 397
2023WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentationcited by 311
2020Multiresolution Knowledge Distillation for Anomaly Detectioncited by 285
2023AnomalyCLIP: Object-agnostic Prompt Learning for Zero-shot Anomaly Detectioncited by 270
2021VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localizationcited by 222
2016A Benchmark for Endoluminal Scene Segmentation of Colonoscopy Imagescited by 220
2024MixLoRA: Enhancing Large Language Models Fine-Tuning with LoRA-based Mixture of Expertscited by 144
2023APRIL-GAN: A Zero-/Few-Shot Anomaly Classification and Segmentation Method for CVPR 2023 VAND Workshop Challenge Tracks 1&2: 1st Place on Zero-shot AD and 4th Place on Few-shot ADcited by 128
2020Deep Learning for Medical Anomaly Detection -- A Surveycited by 127
2024AdaCLIP: Adapting CLIP with Hybrid Learnable Prompts for Zero-Shot Anomaly Detectioncited by 126
2024MambaAD: Exploring State Space Models for Multi-class Unsupervised Anomaly Detectioncited by 113
2024Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detectioncited by 92
2021Deep Learning for Chest X-ray Analysis: A Surveycited by 84
2023BMAD: Benchmarks for Medical Anomaly Detectioncited by 81
2024Asymmetry in Low-Rank Adapters of Foundation Modelscited by 81
2025AA-CLIP: Enhancing Zero-shot Anomaly Detection via Anomaly-Aware CLIPcited by 72
2024A Survey on Visual Anomaly Detection: Challenge, Approach, and Prospectcited by 69
2023CLIP-AD: A Language-Guided Staged Dual-Path Model for Zero-shot Anomaly Detectioncited by 58
2024MuSc: Zero-Shot Industrial Anomaly Classification and Segmentation with Mutual Scoring of the Unlabeled Imagescited by 52
2024FiLo: Zero-Shot Anomaly Detection by Fine-Grained Description and High-Quality Localizationcited by 51
2024VCP-CLIP: A visual context prompting model for zero-shot anomaly segmentationcited by 49
2025Bayesian Prompt Flow Learning for Zero-Shot Anomaly Detectioncited by 31
2023Diversifying the Mixture-of-Experts Representation for Language Models with Orthogonal Optimizercited by 28
2024Learning Unified Reference Representation for Unsupervised Multi-class Anomaly Detectioncited by 19
2025OMoE: Diversifying Mixture of Low-Rank Adaptation by Orthogonal Finetuningcited by 14
2025CoMoE: Contrastive Representation for Mixture-of-Experts in Parameter-Efficient Fine-tuningcited by 4
2025PA-CLIP: Enhancing Zero-Shot Anomaly Detection through Pseudo-Anomaly Awarenesscited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way