metadata
license: cc-by-nc-4.0
task_categories:
- image-classification
- object-detection
- image-segmentation
language:
- en
tags:
- anomaly-detection
- industrial-inspection
- academic-only
pretty_name: ADNet
size_categories:
- 10GB-100GB
ADNet
Access & Licensing
ADNet is released for ACADEMIC RESEARCH PURPOSES ONLY.
The dataset is protected under the CC-BY-NC 4.0 License (Attribution-NonCommercial 4.0 International). Commercial use is strictly prohibited.
To access the dataset, you must apply with your institutional email address (e.g., .edu, .ac.cn).
- Requests from personal emails (e.g., Gmail, Outlook) or commercial domains may be rejected.
- Please clearly state your university/institution and the intended research use in the application form.
Note: We review requests manually. Please allow 1-2 working days for approval.
🔍 Dataset Description
ADNet provides a challenging benchmark for industrial anomaly detection and localization. It covers various object categories with high-resolution images, designed to evaluate the robustness of anomaly detection algorithms in real-world manufacturing scenarios.
- Categories: Multiple industrial objects (e.g., metal parts, textures, electronics).
- Tasks: Image-level Anomaly Detection, Pixel-level Anomaly Localization.
- Format: The dataset is distributed as category-wise
.ziparchives.