First of all

activate virtual environment present inside the env

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

Note: To run webcam_VGG.py make new folder, name it "output", download VGG weights for link below and paste it there

Pre-Trained Model Weights

YOLO Backend weights.

VGG weights for weapon classifcation weights