How jkcamarador/yolo-v4-handgun-detection works
This repository presents a handgun-detection workflow: a collected image dataset is prepared for YOLOv4 training, the trained detector is evaluated, and notebooks demonstrate testing and real-time detection. The README also describes a YOLOv3 benchmark, stratified k-fold validation, and paired significance testing. The tree includes training and testing configurations, class metadata, and links to trained weights. Only generate_train.py was excerpted, so notebook internals and most runtime connections are not verified; links between those stages reflect README statements.
Main entry points: YOLOv4_Handgun_Detection_Training.ipynb YOLOv4 Training · YOLOv4_Handgun_Detection_Testing.ipynb Model Testing · Trained .weights files Gdrive Link.txt Trained Weights · generate_train.py Training Manifest Generator
Components 12
Dataset Preparation
- Gun Image Dataset
- Training Manifest Generator
generate_train.py
YOLOv4 Model
- Training Configuration
yolov4-obj.cfg - Class Metadata
obj.names - Dataset Metadata
obj.data - YOLOv4 Training
YOLOv4_Handgun_Detection_Training.ipynb - Trained Weights
backup/Trained .weights files Gdrive Link.txt
Evaluation and Detection
- Model Testing
YOLOv4_Handgun_Detection_Testing.ipynb - Testing Configuration
yolov4-obj-testing.cfg - Real-time Detection
Realtime_Handgun_Detection_v4.ipynb - YOLOv3 Benchmark
Other components
- Kaggle Dataset
Connections 11
- Kaggle Dataset → Gun Image Dataset: provides images
README.md - Gun Image Dataset → Training Manifest Generator: supplies images
- Training Manifest Generator → YOLOv4 Training: prepares training list
- YOLOv4 Training → Trained Weights: produces weights
README.md - YOLOv4 Training → Training Configuration: uses configuration
- YOLOv4 Training → Class Metadata: uses class labels
- YOLOv4 Training → Dataset Metadata: uses dataset metadata
- Model Testing → Trained Weights: tests model
- Model Testing → Testing Configuration: uses configuration
- Real-time Detection → Trained Weights: runs detector
README.md - YOLOv3 Benchmark → Model Testing: compared with
README.md