mlsanityData to 14 Oct 2026

DevPrompt: Deviation-Based Prompt Learning for One-Normal ShotImage Anomaly Detection

arXiv 2601.15453 ↗First posted 21 Jan 2026cs.CVcited by 0

On the map: Multimodal LLMs / vision-language models → Chart understanding and anomaly detection

Abstract (arXiv)

Few-normal shot anomaly detection (FNSAD) aims to detect abnormal regions in images using only a few normal training samples, making the task highly challenging due to limited supervision and the diversity of potential defects. Recent approaches leverage vision-language models such as CLIP with prompt-based learning to align image and text features. However, existing methods often exhibit weak discriminability between normal and abnormal prompts and lack principled scoring mechanisms for patch-level anomalies. We propose a deviation-guided prompt learning framework that integrates the semantic power of vision-language models with the statistical reliability of deviation-based scoring. Specifically, we replace fixed prompt prefixes with learnable context vectors shared across normal and abnormal prompts, while anomaly-specific suffix tokens enable class-aware alignment. To enhance separability, we introduce a deviation loss with Top-K Multiple Instance Learning (MIL), modeling patch-level features as Gaussian deviations from the normal distribution. This allows the network to assign higher anomaly scores to patches with statistically significant deviations, improving localization and interpretability. Experiments on the MVTecAD and VISA benchmarks demonstrate superior pixel-level detection performance compared to PromptAD and other baselines. Ablation studies further validate the effectiveness of learnable prompts, deviation-based scoring, and the Top-K MIL strategy.

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 17 of its 27 references that are in the map. First posted on arXiv 21 Jan 2026.

Builds on

17 of 27 references.

References · 17

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2016Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzlescited by 1,382
  3. 2021Towards Total Recall in Industrial Anomaly Detectioncited by 776
  4. 2019Memorizing Normality to Detect Anomaly: Memory-augmented Deep Autoencoder for Unsupervised Anomaly Detectioncited by 564
  5. 2018GANomaly: Semi-Supervised Anomaly Detection via Adversarial Trainingcited by 409
  6. 2022SPot-the-Difference Self-Supervised Pre-training for Anomaly Detection and Segmentationcited by 400
  7. 2023WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentationcited by 315
  8. 2019Deep Semi-Supervised Anomaly Detectioncited by 291
  9. 2018Deep Anomaly Detection Using Geometric Transformationscited by 269
  10. 2024PromptAD: Learning Prompts with only Normal Samples for Few-Shot Anomaly Detectioncited by 98
  11. 2024Toward Generalist Anomaly Detection via In-context Residual Learning with Few-shot Sample Promptscited by 89
  12. 2018Learning Representations of Ultrahigh-dimensional Data for Random Distance-based Outlier Detectioncited by 79
5 more
  1. 2021Explainable Deep Few-shot Anomaly Detection with Deviation Networkscited by 69
  2. 2023Unbiased Multiple Instance Learning for Weakly Supervised Video Anomaly Detectioncited by 64
  3. 2019Deep Weakly-supervised Anomaly Detectioncited by 59
  4. 2024Long-Tailed Anomaly Detection with Learnable Class Namescited by 14
  5. 2023Knowledge Distillation for Anomaly Detectioncited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024Bidirectional Multimodal Prompt Learning with Scale-Aware Training for Few-Shot Multi-Class Anomaly Detection
  2. 2026VisualAD: Language-Free Zero-Shot Anomaly Detection via Vision Transformer
  3. 2025Defect-aware Hybrid Prompt Optimization via Progressive Tuning for Zero-Shot Multi-type Anomaly Detection and Segmentation
  4. 2025PromptMoE: Generalizable Zero-Shot Anomaly Detection via Visually-Guided Prompt Mixtures
  5. 2024Do LLMs Understand Visual Anomalies? Uncovering LLM's Capabilities in Zero-shot Anomaly Detection
  6. 2025CoPS: Conditional Prompt Synthesis for Zero-Shot Anomaly Detection
  7. 2025Towards Fine-Grained Vision-Language Alignment for Few-Shot Anomaly Detection
  8. 2024Kernel-Aware Graph Prompt Learning for Few-Shot Anomaly Detection