Introduction #
Authors introduce OD-WeaponDetection: Sohas Detection dataset, featuring a corpus of 5859 images meticulously labeled with 6446 objects spanning six distinct classes, which include knive, pistol, tarjeta, monedero, smartphone, billete categories. This dataset is strategically divided into two key segments: a train split comprising 5002 images and a test split with 857 images. Just as in the example dataset, the authors present this resource as a pivotal contribution to the field of object detection, particularly focusing on weapons and objects of interest within images. The dataset’s rich content is sourced from a myriad of internet-based sources, including frames extracted from YouTube videos and surveillance footage, ensuring it encapsulates real-world challenges. It boasts the inclusion of various knives, mirroring the diverse types, shapes, colors, sizes, and materials that such weapons can embody. Additionally, the dataset addresses complex scenarios where knives are partially obscured by hands and objects simulating their handling, enhancing its practicality. With a diverse array of indoor and outdoor settings, the dataset equips researchers, security professionals, and developers with a versatile resource for advancing object detection models. Detailed information and experimental results are available in a related publication, making the OD-WeaponDetection: Sohos Detection dataset an essential component of the broader Weapon Detection Open Data initiative.
More about Weapon Detection Open Data
The weapon datasets available here are specifically tailored for the development of intelligent video surveillance automatic systems.
An automatic weapon detection system can provide the early detection of potentially violent situations that is of paramount importance for citizens security. One way to prevent these situations is by detecting the presence of dangerous objects such as handguns and knives in surveillance videos. Deep Learning techniques based on Convolutional Neural Networks can be trained to detect this type of object.
The weapon detection task can be performed by different approaches of combining a region proposal technique with a classifier, or integrating both into one model. However, any deep learning model requires to learn a quality image dataset and an annotation according to the classification or detection tasks.
Weapon detection Open Data provides quality image datasets built for training Deep Learning models under the development of an automatic weapon detection system. Weapons datasets for image classification and object detection tasks are described and can be downloaded below. The public datasets are organized depending on the included objects in the dataset images and the target task.
Weapon Detection Open Data structure
Classification
The datasets included in this section have been designed for the classification task based on CNN deep learning models. After the training stage on these datasets, the classification models must distinguish between weapons and different common objects present in the background or handled similarly.
- OD-WeaponDetection: Knife Classification (10 039 images, 100 classes) (available on DatasetNinja)
- OD-WeaponDetection: Pistol Classification (9 857 images, 102 classes) (available on DatasetNinja)
- OD-WeaponDetection: Sohas Classification (9 544 images, 6 classes) (available on DatasetNinja)
Detection
The datasets included in this section have been designed for the object detection task based on Deep Learning architectures with a CNN backbone. The selected images contain weapons and objects but also consider an enriched context of different background objects as well as the way objects are handled. After the training stage on these datasets, the detection models must locate and distinguish between weapons and different common objects present in the background or handled similarly.
- OD-WeaponDetection: Knife Detection (2 078 images, 1 class) (available on DatasetNinja)
- OD-WeaponDetection: Pistol Detection (3 000 images, 1 class) (available on DatasetNinja)
- OD-WeaponDetection: Sohas Detection (5 859 images, 6 classes) (current)
Summary #
OD-WeaponDetection: Sohas Detection is a dataset for an object detection task. It is used in the security industry.
The dataset consists of 5859 images with 6446 labeled objects belonging to 6 different classes including knife, pistol, smartphone, and other: monedero, billete, and tarjeta.
Images in the OD-WeaponDetection: Sohas Detection dataset have bounding box annotations. All images are labeled (i.e. with annotations). There are 2 splits in the dataset: train (5002 images) and test (857 images). The dataset was released in 2020 by the University of Granada, Spain, Ho Chi Minh City University of Technology (HUTECH), Viet Nam, and King Abdulaziz University (KAU) Jeddah, Saudi Arabia.
Here are the visualized examples for the classes:
Explore #
OD-WeaponDetection: Sohas Detection dataset has 5859 images. Click on one of the examples below or open "Explore" tool anytime you need to view dataset images with annotations. This tool has extended visualization capabilities like zoom, translation, objects table, custom filters and more. Hover the mouse over the images to hide or show annotations.
Class balance #
There are 6 annotation classes in the dataset. Find the general statistics and balances for every class in the table below. Click any row to preview images that have labels of the selected class. Sort by column to find the most rare or prevalent classes.
Class ㅤ | Images ㅤ | Objects ㅤ | Count on image average | Area on image average |
---|---|---|---|---|
knife➔ rectangle | 2277 | 2349 | 1.03 | 9.96% |
pistol➔ rectangle | 1510 | 1663 | 1.1 | 45.31% |
smartphone➔ rectangle | 715 | 893 | 1.25 | 29.72% |
monedero➔ rectangle | 611 | 644 | 1.05 | 27.37% |
billete➔ rectangle | 482 | 584 | 1.21 | 25.53% |
tarjeta➔ rectangle | 279 | 313 | 1.12 | 21.31% |
Co-occurrence matrix #
Co-occurrence matrix is an extremely valuable tool that shows you the images for every pair of classes: how many images have objects of both classes at the same time. If you click any cell, you will see those images. We added the tooltip with an explanation for every cell for your convenience, just hover the mouse over a cell to preview the description.
Images #
Explore every single image in the dataset with respect to the number of annotations of each class it has. Click a row to preview selected image. Sort by any column to find anomalies and edge cases. Use horizontal scroll if the table has many columns for a large number of classes in the dataset.
Object distribution #
Interactive heatmap chart for every class with object distribution shows how many images are in the dataset with a certain number of objects of a specific class. Users can click cell and see the list of all corresponding images.
Class sizes #
The table below gives various size properties of objects for every class. Click a row to see the image with annotations of the selected class. Sort columns to find classes with the smallest or largest objects or understand the size differences between classes.
Class | Object count | Avg area | Max area | Min area | Min height | Min height | Max height | Max height | Avg height | Avg height | Min width | Min width | Max width | Max width |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
knife rectangle | 2349 | 9.68% | 88.35% | 0.17% | 23px | 2.26% | 2820px | 99.66% | 196px | 29.65% | 26px | 2.34% | 2744px | 99.95% |
pistol rectangle | 1663 | 41.42% | 99.43% | 0.31% | 27px | 4.91% | 1401px | 99.82% | 254px | 58.08% | 45px | 3.83% | 1808px | 99.82% |
smartphone rectangle | 893 | 24.59% | 99.18% | 0.32% | 52px | 4.81% | 1919px | 99.95% | 715px | 62.57% | 62px | 3.23% | 1916px | 99.79% |
monedero rectangle | 644 | 26.14% | 98.44% | 0.45% | 24px | 6.3% | 1446px | 99.57% | 310px | 43.87% | 79px | 4.64% | 1827px | 99.86% |
billete rectangle | 584 | 21.85% | 99.55% | 0.36% | 43px | 4.72% | 1311px | 99.83% | 288px | 42.43% | 72px | 4.17% | 1816px | 99.9% |
tarjeta rectangle | 313 | 19.41% | 99.09% | 0.12% | 37px | 3.43% | 761px | 99.45% | 217px | 39.46% | 42px | 2.19% | 1069px | 99.86% |
Spatial Heatmap #
The heatmaps below give the spatial distributions of all objects for every class. These visualizations provide insights into the most probable and rare object locations on the image. It helps analyze objects' placements in a dataset.
Objects #
Table contains all 6446 objects. Click a row to preview an image with annotations, and use search or pagination to navigate. Sort columns to find outliers in the dataset.
Object ID ㅤ | Class ㅤ | Image name click row to open | Image size height x width | Height ㅤ | Height ㅤ | Width ㅤ | Width ㅤ | Area ㅤ |
---|---|---|---|---|---|---|---|---|
1➔ | smartphone rectangle | smartphone_1015.jpg | 1281 x 1920 | 454px | 35.44% | 395px | 20.57% | 7.29% |
2➔ | pistol rectangle | pistol_5013.jpg | 1281 x 1920 | 318px | 24.82% | 409px | 21.3% | 5.29% |
3➔ | knife rectangle | DefenseKnifeAttack0376.jpg | 720 x 1280 | 90px | 12.5% | 191px | 14.92% | 1.87% |
4➔ | smartphone rectangle | smartphone_1033.jpg | 1281 x 1920 | 258px | 20.14% | 196px | 10.21% | 2.06% |
5➔ | knife rectangle | LBmframe00160.jpg | 1090 x 1920 | 106px | 9.72% | 61px | 3.18% | 0.31% |
6➔ | knife rectangle | MBsframe00334.jpg | 1090 x 1920 | 70px | 6.42% | 110px | 5.73% | 0.37% |
7➔ | knife rectangle | knife_64.jpg | 529 x 800 | 384px | 72.59% | 634px | 79.25% | 57.53% |
8➔ | knife rectangle | KravMagaKnifeDefenseTechniques062.jpg | 720 x 1280 | 69px | 9.58% | 143px | 11.17% | 1.07% |
9➔ | smartphone rectangle | smartphone_1071.jpg | 1080 x 1920 | 282px | 26.11% | 157px | 8.18% | 2.14% |
10➔ | monedero rectangle | img772.jpg | 1080 x 1920 | 103px | 9.54% | 131px | 6.82% | 0.65% |
License #
OD-WeaponDetection: Sohas Detection is under CC BY-SA 4.0 license.
Citation #
If you make use of the OD-WeaponDetection: Sohas Detection data, please cite the following reference:
@article{article,
author = {Pérez, Francisco and Tabik, Siham and Castillo Lamas, Alberto and Olmos, Roberto and Fujita, Hamido and Herrera, Francisco},
year = {2020},
month = {02},
pages = {105590},
title = {Object Detection Binary Classifiers methodology based on deep learning to identify small objects handled similarly: Application in video surveillance},
volume = {194},
journal = {Knowledge-Based Systems},
doi = {10.1016/j.knosys.2020.105590}
}
If you are happy with Dataset Ninja and use provided visualizations and tools in your work, please cite us:
@misc{ visualization-tools-for-od-weapon-detection-sohas-detection-dataset,
title = { Visualization Tools for OD-WeaponDetection: Sohas Detection Dataset },
type = { Computer Vision Tools },
author = { Dataset Ninja },
howpublished = { \url{ https://datasetninja.com/od-weapon-detection-sohas-detection } },
url = { https://datasetninja.com/od-weapon-detection-sohas-detection },
journal = { Dataset Ninja },
publisher = { Dataset Ninja },
year = { 2024 },
month = { oct },
note = { visited on 2024-10-15 },
}
Download #
Dataset OD-WeaponDetection: Sohas Detection can be downloaded in Supervisely format:
As an alternative, it can be downloaded with dataset-tools package:
pip install --upgrade dataset-tools
… using following python code:
import dataset_tools as dtools
dtools.download(dataset='OD-WeaponDetection: Sohas Detection', dst_dir='~/dataset-ninja/')
Make sure not to overlook the python code example available on the Supervisely Developer Portal. It will give you a clear idea of how to effortlessly work with the downloaded dataset.
The data in original format can be downloaded here.
Disclaimer #
Our gal from the legal dep told us we need to post this:
Dataset Ninja provides visualizations and statistics for some datasets that can be found online and can be downloaded by general audience. Dataset Ninja is not a dataset hosting platform and can only be used for informational purposes. The platform does not claim any rights for the original content, including images, videos, annotations and descriptions. Joint publishing is prohibited.
You take full responsibility when you use datasets presented at Dataset Ninja, as well as other information, including visualizations and statistics we provide. You are in charge of compliance with any dataset license and all other permissions. You are required to navigate datasets homepage and make sure that you can use it. In case of any questions, get in touch with us at hello@datasetninja.com.