AUTOMATIC WEAPON DETECTION &
RECOGNITION USING DEEP LEARNING
Presented By :
[Link] (227R1A6698)
Under the Guidance
of [Link] (227R1A6696)
Dr. S Rao Chintalapudi [Link] (227R1A66B9)
TABLE OF CONTENTS
1. Title Page with Guide Details 8. Sequence Diagram
2. Table of Contents 9. Algorithms/Modules
3. Abstract 10. Dataset Description
4. Disadvantages with Existing System 11. Screen Shots
5. Proposed System 12. Performance Metrics
6. System Architecture 13. Results‑Comparison Graph
7. Usecase Diagram 14. Conclusion and Future Scope
8. Class Diagram 15. Any Queries
ABSTRACT
• Real-time object detection to improve surveillance methods is one of the sought-after applications
of Convolutional Neural Networks (CNNs). This research work has approached the detection of fire
and handguns in areas monitored by cameras.
• Home fires, industrial explosions, and wildfires are a huge problem that cause adverse effects on
the environment. Gun violence and mass shootings are also on the rise in certain parts of the world.
Such incidents are time sensitive and can cause a huge loss to life and property.
• Hence, the proposed work has built a deep learning model based on the YOLOv3 algorithm that
processes a video frame-by-frame to detect such anomalies in real-time and generate an alert for the
concerned authorities.
DISADVANTAGES OF EXISTING SYSTEM
• Poor Performance in Low-Quality Footage.
• Real-Time Processing Challenges.
• Concealed or Partially Visible Weapons
• Cost of Implementation.
• Dependence on Camera Angle and Position.
• High False Positives/Negatives.
• Limited Dataset Diversity.
PROPOSED SYSTEM
• YOLO for Detection – Uses the YOLOv3 object detection system with convolutional
neural networks.
• Fast & Accurate – Provides real-time detection with minimal accuracy loss.
• Darknet53 Architecture – Uses 53 convolutional layers with Leaky ReLU activation and
batch normalization.
• High Accuracy – Efficiently detects fire and guns with good performance
• Deployable on GPU – Works well on GPU-based systems for real-time monitoring.
SYSTEM ARCHITECTURE
USE CASE DIAGRAM
CLASS DIAGRAM
SEQUENCE DIAGRAM
Algorithms/Modules
CNN(Convolutional Neural Network)
• Performs best in accuracy and classification
• Apply filters (kernels) to input images to detect features (e.g., edges, corners)
YOLO (You Only Look Once)
(YOLOv3,YOLOv5,YOLOv7,YOLOv8)
• Very fast, suitable for real-time applications like CCTV
Dataset Description
• Consists of 8,843 images with 11,452 weapon instances.
• Covers five weapon categories:
• Assault rifles
• Handguns
• Knives
• Shotguns
• Sniper rifles
• Images were collected using web scraping with multiple keywords.
Screen Shots • DETECT OUTPUT
PERFORMANCE METRICS
METRICS TYPE ACCURACY
CNN(Convolutional Classifier High
Neural Network)
YOLOv3 One-stage detector ~95%
RESULTS‑COMPARISON GRAPH
CONCLUSION AND FUTURE SCOPE
• Real-Time Fire & Weapon Detection: The project presents an efficient deep learning
model for detecting fire and weapons in real-time using a frame-based approach.
• High Accuracy with YOLO: The model, YOLO, offers strong detection capabilities,
ensuring reliable identification of potential threats.
• Deployment on GPU Systems: The system is optimized for GPU-based deployment,
making it suitable for real-time monitoring applications.
• Enhanced Security & Emergency Response: The technology improves public safety
by reducing response times and assisting law enforcement in mitigating risks.
ANY QUERIES ?
THANK YOU
