Thermal Image Dataset for Concealed Handgun Detection

Description

This dataset provides thermal imaging data designed to support the development and evaluation of detection systems for identifying concealed handguns. It offers a valuable resource for enhancing security technology by enabling the recognition of hidden handguns (also includes a few other objects and could be further extended) through heat signatures. The dataset is licensed under CC BY NC 3.0. However, we do allow commercial use, but only if prior permission is obtained.

Steps to reproduce

The present concealed object detection dataset contains four classes: "Handgun", "Smartphone", "Keys" and "Person". The data were collected using a TOPDON TC001 thermal camera. The dataset is organized as follows: 1. Main Dataset Folder (UCLM_Thermal_Imaging_Dataset): contains two subfolders corresponding to the presence or absence of a handgun in the videos: "Handgun" and "No_Gun". 2. Main Subfolders (e.g., Handgun): these two main subfolders distinguish between situations where the individual carries a concealed handgun and those where he/she carries nothing, a concealed smartphone, or concealed keys. Both the “Handgun” and “No_Gun” folders contain multiple subfolders, one per recorded video (60 in total), each representing a specific scenario. These video subfolders include the following components: - "video.mp4": the video in .mp4 format, showing an individual performing actions related to the category. - "label.json": contains details of the “video.mp4” file, including bounding boxes (in xywh format), frame IDs, and object categories (inside the “No_Gun” folder, videos where no concealed objects are carried, only “Person” annotations are included), along with general information and licenses. At most, two bounding boxes per frame.

Institutions

Categories

Funding

Licence