Skip to content

Latest commit

 

History

History
39 lines (17 loc) · 707 Bytes

File metadata and controls

39 lines (17 loc) · 707 Bytes

Weapon-Detection-in-CCTV-Footage

Repository contains implementation of weapon detection pipeline which consists of 2 independently trained object detection and one classification algorithm.

VGG_16 was trained using PyimageSearch tutorial on gun(~11000) images.

Requirements

Python 3.6 or later

  • numpy
  • opencv-python
  • torch >= 1.1.0

Inference

VGG

images : python3 predict_VGG.py --image test_imgs/gun.jpg

webcame : python3 webcam_VGG.py

Pre-Trained Model Weights

YOLO Backend weights.