Related Experiment Video
Updated: Jul 4, 2025
08:25
Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
Published on: May 7, 2019
9.0K
Firearm-related action recognition and object detection dataset for video surveillance systems
Jesus Ruiz-Santaquiteria1, Juan D Muñoz1, Francisco J Maigler1
1VISILAB, E.T.S. Ingeniería Industrial, University of Castilla-La Mancha, Avda. Camilo José Cela, Ciudad Real, Spain.
Data in Brief
|February 1, 2024
Summary
This new dataset features 398 videos for firearm detection and action recognition in surveillance. It includes meticulously annotated data for handguns and machine guns, aiding machine learning model development for public safety.
Area of Science:
- Computer Vision
- Machine Learning
- Public Safety
Background:
- Developing robust machine learning models for firearm detection in surveillance is crucial.
- Existing datasets may lack the diversity or annotation detail required for comprehensive evaluation.
Purpose of the Study:
- To introduce a novel dataset for firearm detection and action recognition in video surveillance.
- To provide high-quality, annotated data for training and evaluating computer vision models.
Main Methods:
- A dataset of 398 videos was curated, featuring individuals performing surveillance actions.
- Ground truth annotations were meticulously created in COCO JSON format, including bounding boxes.
- Videos were categorized into "Handgun", "Machine_Gun", and "No_Gun".
Main Results:
- The dataset contains detailed annotations essential for precise model evaluation.
- It facilitates research in firearm detection, action recognition, and surveillance applications.
Conclusions:
- This dataset is a valuable resource for advancing research in computer vision for public safety.
- It enables the development and rigorous evaluation of machine learning models for firearm identification in real-world scenarios.
