Fire and Gun Detection

How to use yolo.py:

usage: yolo.py [-h] [--webcam WEBCAM] [--play_video PLAY_VIDEO]
               [--image IMAGE] [--video_path VIDEO_PATH]
               [--image_path IMAGE_PATH] [--verbose VERBOSE]

optional arguments:
  -h, --help            show this help message and exit
  --webcam WEBCAM       True/False
  --play_video PLAY_VIDEO
                        Tue/False
  --image IMAGE         Tue/False
  --video_path VIDEO_PATH
                        Path of video file
  --image_path IMAGE_PATH
                        Path of image to detect objects
  --verbose VERBOSE     To print statements

Weights File Backup

If the GitLFS file is not accessible - Download Weights and keep inside the project folder.

Move inside the project folder and use the following command:

python yolo.py --play_video True --video_path videos/fire1.mp4

Dataset

Training done on google collab - Jupyter notebook

Demo: Youtube

Paper

Fire and Gun Violence based Anomaly Detection System Using Deep Neural Networks
Proceedings of the International conference on Electronics and Sustainable Communication Systems - ICESCS 2020
ISBN: 978-1-7281-4107-7
978-1-7281-4108-4/20/©2020 IEEE