ACD-CLIP: Decoupling Representation and Dynamic Fusion for Zero-Shot Anomaly Detection
On the map: Multimodal LLMs / vision-language models → Chart understanding and anomaly detection
Abstract (arXiv)
Pre-trained Vision-Language Models (VLMs) struggle with Zero-Shot Anomaly Detection (ZSAD) due to a critical adaptation gap: they lack the local inductive biases required for dense prediction and employ inflexible feature fusion paradigms. We address these limitations through an Architectural Co-Design framework that jointly refines feature representation and cross-modal fusion. Our method proposes a parameter-efficient Convolutional Low-Rank Adaptation (Conv-LoRA) adapter to inject local inductive biases for fine-grained representation, and introduces a Dynamic Fusion Gateway (DFG) that leverages visual context to adaptively modulate text prompts, enabling a powerful bidirectional fusion. Extensive experiments on diverse industrial and medical benchmarks demonstrate superior accuracy and robustness, validating that this synergistic co-design is critical for robustly adapting foundation models to dense perception tasks. The source code is available at https://github.com/cockmake/ACD-CLIP.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Chart understanding and anomaly detection, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 21 of its 27 references that are in the map. First posted on arXiv 11 Aug 2025.
Builds on
21 of 27 references.
References · 21
- 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
- 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
- 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
- 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
- 2023Visual Instruction Tuningcited by 7,869
- 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,859
- 2022Flamingo: a Visual Language Model for Few-Shot Learningcited by 4,386
- 2019Parameter-Efficient Transfer Learning for NLPcited by 4,236
- 2016V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentationcited by 2,292
- 2021Towards Total Recall in Industrial Anomaly Detectioncited by 776
- 2020PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localizationcited by 594
- 2019Kvasir-SEG: A Segmented Polyp Datasetcited by 589
9 more
- 2022SPot-the-Difference Self-Supervised Pre-training for Anomaly Detection and Segmentationcited by 400
- 2023WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentationcited by 315
- 2023AnomalyCLIP: Object-agnostic Prompt Learning for Zero-shot Anomaly Detectioncited by 271
- 2021VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localizationcited by 223
- 2024AdaCLIP: Adapting CLIP with Hybrid Learnable Prompts for Zero-Shot Anomaly Detectioncited by 127
- 2022Convolutional Bypasses Are Better Vision Transformer Adapterscited by 121
- 2023BMAD: Benchmarks for Medical Anomaly Detectioncited by 82
- 2024Convolution Meets LoRA: Parameter Efficient Finetuning for Segment Anything Modelcited by 69
- 2023CLIP-AD: A Language-Guided Staged Dual-Path Model for Zero-shot Anomaly Detectioncited by 58
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026VisualAD: Language-Free Zero-Shot Anomaly Detection via Vision Transformer
- 2024Dual-Image Enhanced CLIP for Zero-Shot Anomaly Detection
- 2024Do LLMs Understand Visual Anomalies? Uncovering LLM's Capabilities in Zero-shot Anomaly Detection
- 2025AdaptCLIP: Adapting CLIP for Universal Visual Anomaly Detection
- 2025PromptMoE: Generalizable Zero-Shot Anomaly Detection via Visually-Guided Prompt Mixtures
- 2026WMoE-CLIP: Wavelet-Enhanced Mixture-of-Experts Prompt Learning for Zero-Shot Anomaly Detection
- 2026MoECLIP: Patch-Specialized Experts for Zero-shot Anomaly Detection
- 2025CoPS: Conditional Prompt Synthesis for Zero-Shot Anomaly Detection