dataset_info:
features:
- name: image_id
dtype: int64
- name: image
dtype: image
- name: width
dtype: int64
- name: height
dtype: int64
- name: objects
struct:
- name: bbox
list:
list: float32
length: 4
- name: category
list:
class_label:
names:
'0': GUN
'1': KNIFE
'2': PERSON
splits:
- name: train
num_bytes: 302529532.579
num_examples: 4291
- name: validation
num_bytes: 55386520.228
num_examples: 1029
- name: test
num_bytes: 91500563.04
num_examples: 2295
download_size: 444071476
dataset_size: 449416615.847
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
license: apache-2.0
task_categories:
- image-classification
language:
- en
size_categories:
- 1K<n<10K
Cleaned and Consolidated Weapon Detection Dataset
This dataset is a cleaned, consolidated, and training-ready version of the Subh775/WeaponDetection dataset. The original 29 ambiguous and redundant classes have been merged into 3 distinct, high-level categories: GUN, KNIFE, and PERSON.
This cleaning process makes the dataset significantly more effective for training robust object detection models by removing label ambiguity and creating stronger class definitions. Images that contained no objects belonging to these three final categories have been removed.
Original Source
This dataset is a derivative work based on the weapon-detection Object Detection Model, originally created and shared by yolov7test on Roboflow Universe. The initial Hugging Face version was Subh775/WeaponDetection.
Class Consolidation Details
The original 29 classes were mapped to the new 3 classes as follows:
GUN: [Guns, Guns perspective, Heavy Gun, Long guns, Pistol, Rifle, Shotgun, guns, handgun, heavyweapon, pistol, pistols, rifle, shotgun, weapon]
KNIFE: [Knife, Knife_Deploy, Knife_Weapon, Stabbing]
PERSON: [Person, Aggressor, Hand, Victim, person]
Note: Any classes from the original dataset not listed above (e.g., Blood, violence, larga, etc) have been discarded during the cleaning process.
The dataset is now more robust to the title: Weapon or Threat detection, and can be used to train detection models accordingly.
Citation If you use this dataset in your work, please cite the original creator.
@misc{yolov7test_weapon_detection,
title={Weapon Detection Object Detection Model},
author={yolov7test},
howpublished={\url{[https://universe.roboflow.com/yolov7test-pdxwq/weapon-detection-m7tpo](https://universe.roboflow.com/yolov7test-pdxwq/weapon-detection-m7tpo)}},
year={2022}
}