mlsanityData to 14 Oct 2026

Zero-Shot Learning in Industrial Scenarios: New Large-Scale Benchmark, Challenges and Baseline

arXiv 2606.07965 ↗First posted 6 Jun 2026cs.AIcited by 3

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

Abstract (arXiv)

Large Visual Language Models (LVLMs) have achieved remarkable success in vision tasks. However, the significant differences between industrial and natural scenes make applying LVLMs challenging. Existing LVLMs rely on user-provided prompts to segment objects. This often leads to suboptimal performance due to the inclusion of irrelevant pixels. In addition, the scarcity of data also makes the application of LVLMs in industrial scenarios remain unexplored. To fill this gap, this paper proposes an open industrial dataset and a Refined Text-Visual Prompt (RTVP) for zero-shot industrial defect detection. First, this paper constructs the Multi-Modal Industrial Open Dataset (MMIO) containing 80K+ samples. MMIO contains diverse industrial categories, including 6 super categories and 18 subcategories. MMIO is the first large-scale multi-scenes pre-training dataset for industrial zero-shot learning, and provides valuable training data for open models in future industrial scenarios. Based on MMIO, this paper provides a RTVP specifically for industrial zero-shot tasks. RTVP has two significant advantages: First, this paper designs an expert-guided large model domain adaptation mechanism and designs an industrial zero-shot method based on Mobile-SAM, which enhances the generalization ability of large models in industrial scenarios. Second, RTVP automatically generates visual prompts directly from images and considers text-visual prompt interactions ignored by previous LVLM, improving visual and textual content understanding. RTVP achieves SOTA with 42.2% and 24.7% AP in zero-shot and closed scenes of MMIO.

Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Placed in Chart understanding and anomaly detection, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 27 of its 59 references that are in the map. First posted on arXiv 6 Jun 2026.

Builds on

27 of 59 references.

References · 27

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  3. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
  4. 2017Mask R-CNNcited by 7,277
  5. 2023Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detectioncited by 3,332
  6. 2016Aggregated Residual Transformations for Deep Neural Networkscited by 2,546
  7. 2021YOLOX: Exceeding YOLO Series in 2021cited by 1,071
  8. 2022YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectorscited by 982
  9. 2021Open-vocabulary Object Detection via Vision and Language Knowledge Distillationcited by 906
  10. 2021MDETR -- Modulated Detection for End-to-End Multi-Modal Understandingcited by 819
  11. 2024YOLOv10: Real-Time End-to-End Object Detectioncited by 515
  12. 2024YOLO-World: Real-Time Open-Vocabulary Object Detectioncited by 503
15 more
  1. 2024YOLOv9: Learning What You Want to Learn Using Programmable Gradient Informationcited by 425
  2. 2023Faster Segment Anything: Towards Lightweight SAM for Mobile Applicationscited by 352
  3. 2022YOLOv6: A Single-Stage Object Detection Framework for Industrial Applicationscited by 275
  4. 2023Personalize Segment Anything Model with One Shotcited by 250
  5. 2021VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localizationcited by 223
  6. 2022DetCLIP: Dictionary-Enriched Visual-Concept Paralleled Pre-training for Open-world Detectioncited by 176
  7. 2023Matcher: Segment Anything with One Shot Using All-Purpose Feature Matchingcited by 142
  8. 2020Deep Learning for Medical Anomaly Detection -- A Surveycited by 127
  9. 2022ReCLIP: A Strong Zero-Shot Baseline for Referring Expression Comprehensioncited by 120
  10. 2024VRP-SAM: SAM with Visual Reference Promptcited by 75
  11. 2023How to Efficiently Adapt Large Segmentation Model(SAM) to Medical Imagescited by 53
  12. 2023Application of Segment Anything Model for Civil Infrastructure Defect Assessmentcited by 27
  13. 2023EviPrompt: A Training-Free Evidential Prompt Generation Method for Segment Anything Model in Medical Imagescited by 8
  14. 2026Distilled Large Language Model-Driven Dynamic Sparse Expert Activation Mechanismcited by 7
  15. 2023AdapterShadow: Adapting Segment Anything Model for Shadow Detectioncited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 3.

3 papers.

  1. 2026When Noise Meets Long-Tail: Feature-Threshold Dual Calibration for Robust Pseudo-Labelingcited by 1
  2. 2026Content-Induced Spatial-Spectral Aggregation Network for Change Detection in Remote Sensing Imagescited by 0
  3. 2026Building Change Detection in Earthquake: A Multi-Scale Interaction Network and A Change Detection Datasetcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Unification of Closed-Open Industrial Detection Scenarios: New Large-Scale Benchmarks,Challenges and Baselines
  2. 2024VMAD: Visual-enhanced Multimodal Large Language Model for Zero-Shot Anomaly Detection
  3. 2026SSVP: Synergistic Semantic-Visual Prompting for Industrial Zero-Shot Anomaly Detection
  4. 2024FADE: Few-shot/zero-shot Anomaly Detection Engine using Large Vision-Language Model
  5. 2024MMAD: A Comprehensive Benchmark for Multimodal Large Language Models in Industrial Anomaly Detection
  6. 2025Towards Open-Vocabulary Industrial Defect Understanding with a Large-Scale Multimodal Dataset
  7. 2025Vision-Language In-Context Learning Driven Few-Shot Visual Inspection Model
  8. 2025IADGPT: Unified LVLM for Few-Shot Industrial Anomaly Detection, Localization, and Reasoning via In-Context Learning