DMGDATA: 18,231 Images of Guns, Rifles, and Persons in YOLO Format

DMGDATA is a three-class image dataset for object detection, consisting of guns, rifles, and persons carrying firearms. It contains 18,231 images labeled in the YOLO format and was authored by ayse berika varol malkocoglu. The dataset is scheduled for release on figshare after the associated paper is published.

Use Cases

• Train object detection models to identify firearms based on the three-class labeling.
• Benchmark the performance of YOLO-format models on detecting persons carrying weapons.
• Develop specialized classifiers for security or surveillance scenarios using the provided image categories.

Strengths

• Contains 18,231 annotated images, providing a substantial base for model training.
• Labels are provided in the YOLO format, a standard for object detection tasks.
• Released under a permissive CC-BY-4.0 license.

Limitations

• The dataset is not yet publicly available; access is pending paper publication.
• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

ayse berika varol malkocoglu via figshare

Freshness

Last updated 2026-05-07 17:50:13; freshness should be verified.