Back to datasets
Dataset assetOpen Source CommunityProhibited Item DetectionX‑ray Image Analysis
PIXray Caption dataset
To support research on open‑vocabulary prohibited‑item detection, we contribute the first X‑ray security inspection OVOD benchmark, the PIXray Caption dataset, which contains 15 common categories and 5,046 image‑caption pairs. Sample image‑caption pairs have been released; the full dataset will be made available after paper acceptance.
Source
github
Created
Sep 9, 2024
Updated
Sep 9, 2024
Signals
193 views
Availability
Linked source ready
Overview
Dataset description and usage context
PIXray Caption Dataset
Overview
- Name: PIXray Caption Dataset
- Task: Open‑vocabulary prohibited‑item detection
- Domain: X‑ray security inspection
- Number of Categories: 15
- Number of Image‑Caption Pairs: 5,046
Description
- This dataset is the first benchmark for open‑vocabulary prohibited‑item detection in X‑ray security inspection.
- It includes 5,046 image‑caption pairs across 15 common categories.
- Sample image‑caption pairs have been released; the complete dataset will be released after paper acceptance.
Need downstream help?
Pair the dataset with AI analysis and content workflows.
Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.