The document describes a real-time video-based violence detection system using a convolutional neural network (CNN) to analyze video footage for signs of violent behavior in public areas. The proposed system aims to improve detection accuracy and assist authorities in monitoring violent incidents, addressing the challenges faced by security personnel in identifying perpetrators. By training on the UCF Crime Anomaly Dataset, the system demonstrates effectiveness in detecting and analyzing violent events, contributing to public safety.

1. REAL-TIME VIDEO
BASED VIOLENCE
DETECTION
SYSTEM IN
PUBLIC AREA
USING NEURAL
NETWORK

2. ABSTRACT
Violence-related instances have recently surged
dramatically in areas such remote roads, footpaths,
shopping malls, elevators, sports stadiums, and
liquor stores, which are tragically discovered only
after it is too late. Our project's goal is to develop a
complete system capable of real-time video
analysis, which will aid in detecting the presence of
any violent acts. We achieved an efficient solution
that can be used for real-time analysis of video
footage so that the concerned authority can
monitor the situation. We have put forward a deep
neural network for significant detection accuracy. A
Convolutional Neural Network (CNN) is utilized to
separate frame-level highlights from a video which
are then accumulated utilizing that utilizes
convolutional entryways. CNN and utilized for the
investigation of nearby movement in a video.

3. Aim
The main aim of the project is to
Real-Time Video Based Violence
detection in public using Neural
Network
Objective
Its purpose is to detect signals of
hostility and violence in real-time,
allowing abnormalities to be
distinguished from typical patterns. To
validate our system's performance, it is
trained as well as tested in a large-scale
UCF Crime anomaly dataset

4. 1
Introduction
Violence has increased dramatically in recent years, posing major hazards to people, systems, and
structures. When violence occurs in public, the issue becomes even worse because most people are
not held accountable and cannot be held accountable without proof. Because of their unusual nature,
the majority of horrible crimes occur in public. When we talk about violent activities, we're usually
talking about a strange physical encounter between two or more people. Monitoring the surveillance
has created a lot of difficulty for security personnel because they now have to go through the
footage painstakingly to find the perpetrator and track his movements from one camera to the next,
or view it in real-time to detect violent activities and behavior before or asthey happen.

5. Existing System
•
•
•
•
•
•
Existing System
Drawbacks

6. Proposed
System
• CNN
• Preprocess
• Feature Analysis
• Data Base Loading
• Blob Detection
Are you ready?
Proposed
System
Advantages
• More Accuate results
• Training of Features are easy
Applications • Theft Identification
• Security Analysis

7. START
SYSTEM
ARCHITECTURE
VIDEO
STREAMING
PREPROCESS
BLOB
DETECTION
CNN
TRAINING
VIOLENCE
DETECTION
DATABASE

8. MODULES
• CNN
A Convolutional Neural Network (ConvNet/CNN) is a Deep Learning algorithm that can take in an input image,
assign importance (learnable weights and biases) to various aspects/objects in the image, and be able to
differentiate one from the other
• Blob detection
In computer vision, blob detection methods are aimed at detecting regions in a digital image that differ in
properties, such as brightness or color, compared to surrounding regions.
• Pre-Processing
Pre-processing is a common name for operations with images at the lowest level of abstraction both input and
output are intensity images.The aim of pre-processing is an improvement of the image data that suppresses
unwanted distortions or enhances some image features important for further processing
STOP

9. System Requirements
SOFTWARE
REQUIREMENT
HARDWARE
REQUIREMENT
Python / Anaconda Navigator
Google Colab
Python Language
Jupyter Notebook
OS – Windows 7, 8, and 10
(32 and 64-bit)
.RAM – 4 GB

10. What is computer vision?
• In Object Classification, You Train A Model On A Dataset Of Specific Objects, And The
Model Classifies New Objects As Belonging To One Or More Of Your Training Categories.
• For Object Identification, Your Model Will Recognize A Specific Instance Of An Object –
For Example, Parsing Two Faces In An Image And Tagging One As Tom Cruise And One
As Katie Holmes.

11. Business Use Cases For Computer Vision
Computer Vision Is One Of The Areas In Machine Learning Where Core Concepts Are
Already Being Integrated Into Major Products That We Use Every Day. Google Is Using Maps
To Leverage Their Image Data And Identify Street Names, Businesses, And Office Buildings.
Facebook Is Using Computer Vision To Identify People In Photos And Do A Number Of
Things With That Information.

12. Computer Vision On Algorithmia
• Salnet Automatically Identifies The Most Important Parts Of An Image
• Nudity Detection Detects Nudity In Pictures
• Emotion Recognition Parses Emotions Exhibited In Images
• Deepstyle Transfers Next-Level Filters Onto Your Image
• Face Recognition…Recognizes Faces.
• Image Memorability Judges How Memorable An Image Is.

13. Packages and Frameworks
• Opencv – “Opencv Was Designed For Computational Efficiency And With A Strong
Focus On Real-Time Applications. Adopted All Around The World, Opencv Has More
Than 47 Thousand People Of User Community And Estimated Number Of Downloads
Exceeding 14 Million. Usage Ranges From Interactive Art, To Mines Inspection, Stitching
Maps On The Web Or Through Advanced Robotics.”
• Simplecv – “Simplecv Is An Open Source Framework For Building Computer Vision
Applications. With It, You Get Access To Several High-Powered Computer Vision
Libraries Such As Opencv – Without Having To First Learn About Bit Depths, File
Formats, Color Spaces, Buffer Management, Eigenvalues, Or Matrix Versus Bitmap
Storage.”

14. TensorFlow
• the most famous deep learning library in the world is Google's TensorFlow. Google
product uses machine learning in all of its products to improve the search engine,
translation, image captioning or recommendations.
• To give a concrete example, Google users can experience a faster and more refined the
search with AI. If the user types a keyword a the search bar, Google provides a
recommendation about what could be the next word.

15. TensorFlow Architecture
Tensor flow architecture works in three parts:
• Pre-processing the data
• Build the model
• Train and estimate the model

16. Where can Tensor flow run?
• TensorFlow can hardware, and software requirements can be classified into
• Development Phase: This is when you train the mode. Training is usually done on your
Desktop or laptop.
• Run Phase or Inference Phase: Once training is done Tensorflow can be run on many
different platforms. You can run it on
• Desktop running Windows, macOS or Linux
• Cloud as a web service
• Mobile devices like iOS and Android

17. Hardware Requirements
• Operating System: Windows Only
• Processor : i5 and above
• Ram: 4GB and above
• Hard Disk: 50 GB

18. Functional Requirements
• Data Collection
• Data Preprocessing
• Training And Testing
• Modeling
• Predicting

19. Results

20. Results

21. Conclusion
The main objective of this study is to detect fight scenes from surveillance cameras in a fast and accurate way.
Nowadays, the rate of violence in our environment is escalating, posing a hazard to people, structures, and
systems. There has always been a need for a better system to assist the police in monitoring the violence, which is
often difficult to handle because it is a group activity and the elimination process to discover the perpetrator is
time-consuming. The results of our experiment show that CNN architecture may be used effectively to train a
model over Hockey's dataset. Our work can assist law enforcement in keeping their area under control. By
consolidating CNN with the accuracy increases to a specific edge when contrasted with transfer learning models
alone.

22. References
[1] E. B. Nievas, O. D. Suarez, G. B. Garc´ıa, and R. Sukthankar, “Violence detection in video using computer vision
techniques,” in the International conference on Computer analysis of images and patterns. Springer, 2011, pp. 332–
339.
[2] C. Ding, S. Fan, M. Zhu, W. Feng, and B. Jia, “Violence detection in video by using 3d convolutional neural
networks,” in International Symposium on Visual Computing. Springer,2014, pp. 551–558.
[3] Z. Dong, J. Qin, and Y. Wang, “Multi-stream deep networks for person to person violence detection in videos,” in
Chinese Conference on Pattern Recognition. Springer, 2016, pp. 517– 531.
[4] Sudhakaran and O. Lanz, “Learning to detect violent videos using convolutional long short?term memory,”
published in 2017 14th IEEE International Conference on Advanced Video andSignal Based Surveillance (AVSS). IEEE,
2017, pp. 1–6.
[5] Hanson, K. Pnvr, S. Krishnagopal, and L. Davis, “Bidirectional Convolutional LSTM forthe detection of violence in
videos,” in Proceedings of the European Conference on Computer Vision (ECCV), 2018, pp. 0–0. Dai.
[6] J. Li, X. Jiang, T. Sun, and K. Xu, “Efficient violence detection using 3d convolutional neural networks,” in 2019 16th
IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS). IEEE, 2019, pp. 1–8

23. Thank you
— ANY QUESTIONS