Weapon Detection Using Neural Networks in Real-Time Surveillance Video

A Practical Implementation of the Faster R-CNN Algorithm for Object Detection; detecting guns, knives and non-weapon objects from the webcam video stream:

• Setting up deep learning in Windows OS with TensorFlow.
• Processing datasets for supervised training.
• Using Google Colaboratory for cloud training.
• Applied the fundamental mechanism of machine learning for supervised learning

Background

This project was created for my final year project. It's about using Faster R-CNN model for detecting guns, knives and non-weapon objects (can be any non-weaponry object) in a real-time video stream. The purpose of this project was not aim for high accuracy, it was for me to study the result and effect of non-weapon objects on the accuracy. In addition, this hands-on project helped me to understand the fundamental knowledge of machine learning and computer vision.

For further information and details about how I created this project, please refer to the following report written by me:

Weapon Detection Using Neural Networks in Real-Time Surveillance Video