SlideShare uma empresa Scribd logo
1 de 37
Baixar para ler offline
Gippsland School of Information Technology (GSIT)

Behaviour Recognition
Framework for Intelligent
Visual Surveillance
06 April 2009
www.monash.edu.au
Project Team
• Mahfuzul Haque, PhD student (2 yrs, 1 m)
• A/Prof. Manzur Murshed
• Dr. Manoranjan Paul

www.monash.edu.au
2
Project Motivation
“Behaviour Recognition Framework for
Intelligent Visual Surveillance”

Why “Intelligent” Surveillance?
Why “Behaviour Recognition”?
What type of “Behaviours”?

www.monash.edu.au
3
Surveillance Everywhere

Are we really protected?

www.monash.edu.au
4
Too Many Cameras

Deployment of large number of surveillance cameras in recent years
London Heathrow airport has more than 5000 cameras!!
www.monash.edu.au
5
Behind the Scene: Worried Human Monitor

Dependability on human monitors has increased.
Reliability on surveillance system has decreased.
www.monash.edu.au
6
Project Goals
 Aiding human monitors by
automatic detection of
specific abnormal behaviors
 Decreasing dependability on
human monitors
 Improving reliability of
surveillance systems for
ensuring human security
www.monash.edu.au
7
Project Scope
Group Behaviours







Mob Violence
Crowding
Sudden Group Formation
Sudden Group Deformation
Shooting
Panic Driven Behaviours

www.monash.edu.au
8
Research Question

How to recognize specific group
behaviours from surveillance video
streams in real-time?
Research Area
- Computer Vision
- Application of Machine Learning
www.monash.edu.au
9
System Architecture

Behaviour Profile

Surveillance
Video Stream
Behaviour Recognition
Framework

www.monash.edu.au
10
System Architecture

Behaviour Profile

Surveillance
Video Stream
Behaviour Recognition
Framework

www.monash.edu.au
11
Behaviour Profile
Surveillance Video Stream (System Input)
Time

0

Unknown

20

Group
Appearing

60

Group
Appearing

140

320

Group
Merging

Group
Splitting

Behaviour Profile (System Output)
www.monash.edu.au
12
System Architecture

Behaviour Profile

Surveillance
Video Stream
Behaviour Recognition
Framework

www.monash.edu.au
13
Behaviour Recognition Framework

Framework Components
•
•
•
•

Background Modelling
Frame Level Feature Extraction
Temporal Feature Extraction
Behaviour Classification

www.monash.edu.au
14
Behaviour Recognition Framework

Background
Modelling

Frame Level
Feature
Extraction

Temporal
Feature
Extraction

Behaviour
Classification

www.monash.edu.au
15
Background Modelling

Background
Modelling

Frame Level
Feature
Extraction

Temporal
Feature
Extraction

Behaviour
Classification

www.monash.edu.au
16
Background Modelling
How to extract the active regions from surveillance video stream?

Background Subtraction
Current frame

Challenges!!
=

Background

Moving foreground

www.monash.edu.au
17
Background Modelling
σ2

P(x)

µ
P(x)

x

Sky
Cloud
Leaf
Moving Person

σ2

Road
Shadow
Moving Car

Floor
Shadow
Walking People

Cloud
µ
P(x)

x

P(x)

Person
Leaf
Sky

σ2

µ

x

x (Pixel intensity)

www.monash.edu.au
18
Background Modelling
Background
Model
Current frame

Moving foreground

Frame 1

Frame N

Background Models
road

shadow

car

shadow

road
Models are ordered by ω/σ

ω1
σ12
µ1
road

ω2
σ22
µ2
shadow

65%

20%

ω3
σ32
µ3
car

15%
www.monash.edu.au
19
Background Modelling
First
Frame

Test
Frame

Ground
Truth

S&G

Lee

Proposed

(1)

(2)

(3)

(4)

(5)

(1) PETS2000; (2) PETS2006-B1; (3) PETS2006-B2; (4) PETS2006-B3; and (5) PETS2006-B4.

www.monash.edu.au
20
Frame Level Feature Extraction

Background
Modelling

Frame Level
Feature
Extraction

Temporal
Feature
Extraction

Behaviour
Classification

www.monash.edu.au
21
Frame Level Feature Extraction
• Feature Categories:
– Count
– Area
– Density
– Bounding Box
– Filling Ratio
– Aspect Ratio
• 30 frame level features
Bounding Boxes
www.monash.edu.au
22
Frame Level Feature Extraction
Foreground Count

Foreground Area

Foreground Density

• FC (Foreground Count)

• TFA (Total Foreground Area)
• AFA (Average Foreground Area)
• VFA (Variance of Foreground Area)
• MAXFA (Maximum Foreground Area)
• MINFA (Minimum Foreground Area)

• AFD (Average Foreground Density)
• VFD (Variance of Foreground Density)

Filling Ratio

Bounding Box – Area

Bounding Box – Width

• TFR (Total Filling Ratio)
• AFR (Average Filling Ratio)
• VFR (Variance of Filling Ratio)
• MAXFR (Maximum Filling Ratio)
• MINFR (Minimum Filling Ratio)

• TBBA (Total Bounding Box Area)
• ABBA (Average Bounding Box Area)
• VBBA (Variance of Bounding Box Area)
• MAXBBA (Maximum Bounding Box
Area)
• MINBBA (Minimum Bounding Box Area)

• ABBW (Average Bounding Box Width)
• VBBW (Variance of Bounding Box
Width)
• MAXBBW (Maximum Bounding Box
Width)
• MINBBW (Minimum Bounding Box
Width)

Bounding Box – Height

Aspect Ratio

• ABBH (Average Bounding Box Height)
• VBBH (Variance of Bounding Box
Height)
• MAXBBH (Maximum Bounding Box
Height)
• MINBBH (Minimum Bounding Box
Height)

• AAR (Average Aspect Ratio)
• VAR (Variance of Aspect Ratio)
• MAXAR (Maximum Aspect Ratio)
• MINAR (Minimum Aspect Ratio)

www.monash.edu.au
23
Temporal Feature Extraction

Background
Modelling

Frame Level
Feature
Extraction

Temporal
Feature
Extraction

Behaviour
Classification

www.monash.edu.au
24
Temporal Feature Extraction
• Fixed length, partially overlapped
sliding window
• Temporal data smoothing – polynomial
curve fitting
• 9 temporal features for each frame level
feature
• Output: 270 temporal features

www.monash.edu.au
25
Temporal Feature Extraction
TFA (Total Foreground Area)

Temporal Features
TFA (%)

• MAX
• MIN
• AVG
• VAR
• RATE
• TIME(MAX)
• TIME(MIN)
• D = TIME(MAX) - TIME(MIN)
• SLOPE ( D/2 )

Time (window = 100 frames)
www.monash.edu.au
26
Temporal Feature Extraction
TFA (Total Foreground Area)

Temporal Features

TFA (%)

MAX

MIN

• MAX
• MIN
• AVG
• VAR
• RATE
• TIME(MAX)
• TIME(MIN)
• D = TIME(MAX) - TIME(MIN)
• SLOPE ( D/2 )

TIME(MAX)
TIME(MIN)

Time (window = 100 frames)
www.monash.edu.au
27
Behaviour Classification

Background
Modelling

Frame Level
Feature
Extraction

Temporal
Feature
Extraction

Behaviour
Classification

www.monash.edu.au
28
Behaviour Classification
•
•
•
•
•
•
•
•

Individual classifiers for each behaviour
Supervised training
Feature ranking
Top 100 features from 270 features
Dimension reduction (PCA)
Max dimension 30
SVM classifier
Output: Behaviour Profile
www.monash.edu.au
29
Behaviour Classification
Experiments
GROUP FORMING
• Accuracy: 0.9767
• Top 3 features
• TIME(MAX)-VFD
• TIME(MAX)-AFD
• TIME(MAX) - TIME(MIN)-VFD

GROUP SPLITTING AND SPREADING
• Accuracy: 0.8488
• Top 3 features
• TIME(MAX)-VFD
• TIME(MIN)-ABBA
• TIME(MIN)-AFA

BLOCKED EXIT
• Accuracy: 0.9651
• Top 3 features
• TIME(MIN)-TFA
• MIN-MINAR
• TIME(MAX)-TFA

www.monash.edu.au
30
Summary: Framework Components

Background
Modelling

Frame Level Feature
Extraction

Temporal Feature
Extraction

Behaviour
Classification

• Multiple Background
Models
• Gaussian Mixture
Models (GMM)
• Unsupervised
• Output: Foreground
Region/Mask

• Feature Categories:
• Count
• Area
• Density
• Bounding Box
• Filling Ratio
• Aspect Ratio
• Output: 30 Frame Level
Features

• Fixed Length, Partially
Overlapped Sliding
Window
• Temporal Data
Smoothing –
Polynomial Curve
Fitting
• 9 Temporal Features for
Each Frame Level
Features
• Output: 270 Temporal
Features

• Individual Classifiers for
Each Behaviour
• Each Classifier is
Trained Using
Supervised Learning
• Feature Ranking
• Top 100 Features
• Dimension Reduction
(PCA)
• Max Dimension 30
• SVM classifier
• Output: Behaviour
Profile

www.monash.edu.au
31
Summary: Framework Components

Background
Modelling

Frame Level Feature
Extraction

Temporal Feature
Extraction

Behaviour
Classification

• Multiple Background
Models
• Gaussian Mixture
Models (GMM)
• Unsupervised
• Output: Foreground
Region/Mask

• Feature Categories:
• Count
• Area
• Density
• Bounding Box
• Filling Ratio
• Aspect Ratio
• Output: 30 Frame
Level Features

• Fixed Length, Partially
Overlapped Sliding
Window
• Temporal Data
Smoothing –
Polynomial Curve
Fitting
• 9 Temporal Features for
Each Frame Level
Features
• Output: 270 Temporal
Features

• Individual Classifiers for
Each Behaviour
• Each Classifier is
Trained Using
Supervised Learning
• Feature Ranking
• Top 100 Features
• Dimension Reduction
(PCA)
• Max Dimension 30
• SVM classifier
• Output: Behaviour
Profile

www.monash.edu.au
32
Research Challenges
• No tracking/trajectory
• Simple behaviours
– Group appear/disappear
– Group merge/split

• Panic driven behaviours
– Fire/Blocked exit
– Fighting/Shooting

• Context variation
– Speed
– Direction
– Object Resolution
www.monash.edu.au
33
Implemented System: VSTK

www.monash.edu.au
34
Publications
1. Mahfuzul Haque, Manzur Murshed, and Manoranjan Paul, Improved
Gaussian Mixtures for Robust Object Detection by Adaptive MultiBackground Generation, International Conference on Pattern
Recognition (ICPR), Tampa, Florida, USA, 2008. (CORE A)
2. Mahfuzul Haque, Manzur Murshed, and Manoranjan Paul, A Hybrid
Object Detection Technique from Dynamic Background Using
Gaussian Mixture Models, IEEE International Workshop on Multimedia
Signal Processing (MMSP), Cairns, Australia, 2008. (CORE A)
3. Mahfuzul Haque, Manzur Murshed, and Manoranjan Paul, On Stable
Dynamic Background Generation Technique using Gaussian Mixture
Models for Robust Object Detection, IEEE International Conference On
Advanced Video and Signal Based Surveillance (AVSS), Santa Fe,
New Mexico, USA, 2008. (CORE A)
CORE - COmputing Research and Education Association
www.monash.edu.au
35
Thank you!

Q&A
Mahfuzul.Haque@infotech.monash.edu.au
http://www.mahfuzulhaque.com

www.monash.edu.au
36
Acknowledgments
URLs of the images used in this presentation
•
•
•
•
•
•
•
•
•
•
•
•
•
•

http://www.fotosearch.com/DGV464/766029/
http://www.cyprus-trader.com/images/alert.gif
http://security.polito.it/~lioy/img/einstein8ci.jpg
http://www.dtsc.ca.gov/PollutionPrevention/images/question.jpg
http://www.unmikonline.org/civpol/photos/thematic/violence/streetvio2.jpg
http://www.airports-worldwide.com/img/uk/heathrow00.jpg
http://www.highprogrammer.com/alan/gaming/cons/trips/genconindy2003/exhibithall-crowd-2.jpg
http://www.bhopal.org/fcunited/archives/fcu-crowd.jpg
http://img.dailymail.co.uk/i/pix/2006/08/passaPA_450x300.jpg
http://www.defenestrator.org/drp/files/surveillance-cameras-400.jpg
http://www.cityofsound.com/photos/centre_poin/crowd.jpg
http://www.hindu.com/2007/08/31/images/2007083156401501.jpg
http://paulaoffutt.com/pics/images/crowd-surfing.jpg
http://msnbcmedia1.msn.com/j/msnbc/Components/Photos/070225/070225_surv
eillance_hmed.hmedium.jpg

www.monash.edu.au
37

Mais conteúdo relacionado

Semelhante a Talk 2009-monash-seminar-intelligent-video-surveillance

Talk 2009-monash-seminar-perception
Talk 2009-monash-seminar-perceptionTalk 2009-monash-seminar-perception
Talk 2009-monash-seminar-perception
Mahfuzul Haque
 
Load Balancing In Cloud Computing newppt
Load Balancing In Cloud Computing newpptLoad Balancing In Cloud Computing newppt
Load Balancing In Cloud Computing newppt
Utshab Saha
 
Talk 2010-monash-seminar-panic-driven-event-detection
Talk 2010-monash-seminar-panic-driven-event-detectionTalk 2010-monash-seminar-panic-driven-event-detection
Talk 2010-monash-seminar-panic-driven-event-detection
Mahfuzul Haque
 
Program on Mathematical and Statistical Methods for Climate and the Earth Sys...
Program on Mathematical and Statistical Methods for Climate and the Earth Sys...Program on Mathematical and Statistical Methods for Climate and the Earth Sys...
Program on Mathematical and Statistical Methods for Climate and the Earth Sys...
The Statistical and Applied Mathematical Sciences Institute
 
Sampling-SDM2012_Jun
Sampling-SDM2012_JunSampling-SDM2012_Jun
Sampling-SDM2012_Jun
MDO_Lab
 
Development and Initial Testing of an Autonomous Surface
Development and Initial Testing of an Autonomous SurfaceDevelopment and Initial Testing of an Autonomous Surface
Development and Initial Testing of an Autonomous Surface
Mohamad Hilmi Mat Idris
 

Semelhante a Talk 2009-monash-seminar-intelligent-video-surveillance (20)

Talk 2009-monash-seminar-perception
Talk 2009-monash-seminar-perceptionTalk 2009-monash-seminar-perception
Talk 2009-monash-seminar-perception
 
Study of Tactile interactions for visually disabled and hearing impaired
Study of Tactile interactions for visually disabled and hearing impaired Study of Tactile interactions for visually disabled and hearing impaired
Study of Tactile interactions for visually disabled and hearing impaired
 
Load Balancing In Cloud Computing newppt
Load Balancing In Cloud Computing newpptLoad Balancing In Cloud Computing newppt
Load Balancing In Cloud Computing newppt
 
Talk 2010-monash-seminar-panic-driven-event-detection
Talk 2010-monash-seminar-panic-driven-event-detectionTalk 2010-monash-seminar-panic-driven-event-detection
Talk 2010-monash-seminar-panic-driven-event-detection
 
TraitCapture:Open source tools for DIY high throughput Phenomics and NextGen ...
TraitCapture:Open source tools for DIY high throughput Phenomics and NextGen ...TraitCapture:Open source tools for DIY high throughput Phenomics and NextGen ...
TraitCapture:Open source tools for DIY high throughput Phenomics and NextGen ...
 
YARN Federation
YARN Federation YARN Federation
YARN Federation
 
1196 dubowsky[1]
1196 dubowsky[1]1196 dubowsky[1]
1196 dubowsky[1]
 
Monitoring tropical forest cover Activities of ONFI in remote sensing
Monitoring tropical forest cover Activities of ONFI in remote sensingMonitoring tropical forest cover Activities of ONFI in remote sensing
Monitoring tropical forest cover Activities of ONFI in remote sensing
 
Program on Mathematical and Statistical Methods for Climate and the Earth Sys...
Program on Mathematical and Statistical Methods for Climate and the Earth Sys...Program on Mathematical and Statistical Methods for Climate and the Earth Sys...
Program on Mathematical and Statistical Methods for Climate and the Earth Sys...
 
Sampling-SDM2012_Jun
Sampling-SDM2012_JunSampling-SDM2012_Jun
Sampling-SDM2012_Jun
 
Session 2 intercropping dst
Session 2 intercropping dstSession 2 intercropping dst
Session 2 intercropping dst
 
Development and Initial Testing of an Autonomous Surface
Development and Initial Testing of an Autonomous SurfaceDevelopment and Initial Testing of an Autonomous Surface
Development and Initial Testing of an Autonomous Surface
 
PhD course in amplicon sequencing at midas workshop 2019
PhD course in amplicon sequencing at midas workshop 2019PhD course in amplicon sequencing at midas workshop 2019
PhD course in amplicon sequencing at midas workshop 2019
 
An introduction to variable and feature selection
An introduction to variable and feature selectionAn introduction to variable and feature selection
An introduction to variable and feature selection
 
NIDM-Results. A standard for describing and sharing neuroimaging results: app...
NIDM-Results. A standard for describing and sharing neuroimaging results: app...NIDM-Results. A standard for describing and sharing neuroimaging results: app...
NIDM-Results. A standard for describing and sharing neuroimaging results: app...
 
Ausplots Training - Session 5
Ausplots Training - Session 5Ausplots Training - Session 5
Ausplots Training - Session 5
 
YOLOv4: optimal speed and accuracy of object detection review
YOLOv4: optimal speed and accuracy of object detection reviewYOLOv4: optimal speed and accuracy of object detection review
YOLOv4: optimal speed and accuracy of object detection review
 
Parameters for drive test
Parameters for drive testParameters for drive test
Parameters for drive test
 
From pixels to point clouds - Using drones,game engines and virtual reality t...
From pixels to point clouds - Using drones,game engines and virtual reality t...From pixels to point clouds - Using drones,game engines and virtual reality t...
From pixels to point clouds - Using drones,game engines and virtual reality t...
 
陸永祥/全球網路攝影機帶來的機會與挑戰
陸永祥/全球網路攝影機帶來的機會與挑戰陸永祥/全球網路攝影機帶來的機會與挑戰
陸永祥/全球網路攝影機帶來的機會與挑戰
 

Mais de Mahfuzul Haque

Talk 2011-buet-perception-event
Talk 2011-buet-perception-eventTalk 2011-buet-perception-event
Talk 2011-buet-perception-event
Mahfuzul Haque
 
Talk 2009-monash-open-day-surveillance
Talk 2009-monash-open-day-surveillanceTalk 2009-monash-open-day-surveillance
Talk 2009-monash-open-day-surveillance
Mahfuzul Haque
 
Talk 2007-monash-seminar-behavior-recognition-framework
Talk 2007-monash-seminar-behavior-recognition-frameworkTalk 2007-monash-seminar-behavior-recognition-framework
Talk 2007-monash-seminar-behavior-recognition-framework
Mahfuzul Haque
 
Kb behaviour-recognition
Kb behaviour-recognitionKb behaviour-recognition
Kb behaviour-recognition
Mahfuzul Haque
 
Talk 2012-icmew-perception
Talk 2012-icmew-perceptionTalk 2012-icmew-perception
Talk 2012-icmew-perception
Mahfuzul Haque
 
Poster: Monash Research Month 2009
Poster: Monash Research Month 2009Poster: Monash Research Month 2009
Poster: Monash Research Month 2009
Mahfuzul Haque
 
Poster: Monash Research Month 2008
Poster: Monash Research Month 2008Poster: Monash Research Month 2008
Poster: Monash Research Month 2008
Mahfuzul Haque
 
Poster: Monash Research Month 2007
Poster: Monash Research Month 2007Poster: Monash Research Month 2007
Poster: Monash Research Month 2007
Mahfuzul Haque
 
Poster: EII Workshop 2007
Poster: EII Workshop 2007Poster: EII Workshop 2007
Poster: EII Workshop 2007
Mahfuzul Haque
 
Poster: EII Winter School 2007
Poster: EII Winter School 2007Poster: EII Winter School 2007
Poster: EII Winter School 2007
Mahfuzul Haque
 

Mais de Mahfuzul Haque (19)

Dependency inversion using ports and adapters
Dependency inversion using ports and adaptersDependency inversion using ports and adapters
Dependency inversion using ports and adapters
 
Resilient machine learning systems for health analytics
Resilient machine learning systems for health analyticsResilient machine learning systems for health analytics
Resilient machine learning systems for health analytics
 
Talk 2012-icmew-event
Talk 2012-icmew-eventTalk 2012-icmew-event
Talk 2012-icmew-event
 
Talk 2011-buet-perception-event
Talk 2011-buet-perception-eventTalk 2011-buet-perception-event
Talk 2011-buet-perception-event
 
Talk 2009-monash-open-day-surveillance
Talk 2009-monash-open-day-surveillanceTalk 2009-monash-open-day-surveillance
Talk 2009-monash-open-day-surveillance
 
Talk 2007-monash-seminar-behavior-recognition-framework
Talk 2007-monash-seminar-behavior-recognition-frameworkTalk 2007-monash-seminar-behavior-recognition-framework
Talk 2007-monash-seminar-behavior-recognition-framework
 
Kb hmm
Kb hmmKb hmm
Kb hmm
 
Kb gait-recognition
Kb gait-recognitionKb gait-recognition
Kb gait-recognition
 
Kb behaviour-recognition
Kb behaviour-recognitionKb behaviour-recognition
Kb behaviour-recognition
 
Talk 2012-icmew-perception
Talk 2012-icmew-perceptionTalk 2012-icmew-perception
Talk 2012-icmew-perception
 
Poster: Monash Research Month 2009
Poster: Monash Research Month 2009Poster: Monash Research Month 2009
Poster: Monash Research Month 2009
 
Poster: Monash Research Month 2008
Poster: Monash Research Month 2008Poster: Monash Research Month 2008
Poster: Monash Research Month 2008
 
Poster: Monash Research Month 2007
Poster: Monash Research Month 2007Poster: Monash Research Month 2007
Poster: Monash Research Month 2007
 
Poster: ICPR 2008
Poster: ICPR 2008Poster: ICPR 2008
Poster: ICPR 2008
 
Poster: ICME 2010
Poster: ICME 2010Poster: ICME 2010
Poster: ICME 2010
 
Poster: EII Workshop 2007
Poster: EII Workshop 2007Poster: EII Workshop 2007
Poster: EII Workshop 2007
 
Poster: EII Winter School 2007
Poster: EII Winter School 2007Poster: EII Winter School 2007
Poster: EII Winter School 2007
 
Poster: AVSS 2012
Poster: AVSS 2012Poster: AVSS 2012
Poster: AVSS 2012
 
Poster: MMSP 2008
Poster: MMSP 2008Poster: MMSP 2008
Poster: MMSP 2008
 

Último

Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers:  A Deep Dive into Serverless Spatial Data and FMECloud Frontiers:  A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Safe Software
 
Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire business
panagenda
 

Último (20)

Apidays Singapore 2024 - Scalable LLM APIs for AI and Generative AI Applicati...
Apidays Singapore 2024 - Scalable LLM APIs for AI and Generative AI Applicati...Apidays Singapore 2024 - Scalable LLM APIs for AI and Generative AI Applicati...
Apidays Singapore 2024 - Scalable LLM APIs for AI and Generative AI Applicati...
 
EMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWER
EMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWEREMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWER
EMPOWERMENT TECHNOLOGY GRADE 11 QUARTER 2 REVIEWER
 
A Beginners Guide to Building a RAG App Using Open Source Milvus
A Beginners Guide to Building a RAG App Using Open Source MilvusA Beginners Guide to Building a RAG App Using Open Source Milvus
A Beginners Guide to Building a RAG App Using Open Source Milvus
 
Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)
 
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
 
"I see eyes in my soup": How Delivery Hero implemented the safety system for ...
"I see eyes in my soup": How Delivery Hero implemented the safety system for ..."I see eyes in my soup": How Delivery Hero implemented the safety system for ...
"I see eyes in my soup": How Delivery Hero implemented the safety system for ...
 
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers:  A Deep Dive into Serverless Spatial Data and FMECloud Frontiers:  A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
 
Artificial Intelligence Chap.5 : Uncertainty
Artificial Intelligence Chap.5 : UncertaintyArtificial Intelligence Chap.5 : Uncertainty
Artificial Intelligence Chap.5 : Uncertainty
 
Corporate and higher education May webinar.pptx
Corporate and higher education May webinar.pptxCorporate and higher education May webinar.pptx
Corporate and higher education May webinar.pptx
 
MS Copilot expands with MS Graph connectors
MS Copilot expands with MS Graph connectorsMS Copilot expands with MS Graph connectors
MS Copilot expands with MS Graph connectors
 
Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire business
 
TrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data DiscoveryTrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
 
Manulife - Insurer Transformation Award 2024
Manulife - Insurer Transformation Award 2024Manulife - Insurer Transformation Award 2024
Manulife - Insurer Transformation Award 2024
 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
 
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
 
AWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of TerraformAWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of Terraform
 
Web Form Automation for Bonterra Impact Management (fka Social Solutions Apri...
Web Form Automation for Bonterra Impact Management (fka Social Solutions Apri...Web Form Automation for Bonterra Impact Management (fka Social Solutions Apri...
Web Form Automation for Bonterra Impact Management (fka Social Solutions Apri...
 
MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024
 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
 
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbu
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu SubbuApidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbu
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbu
 

Talk 2009-monash-seminar-intelligent-video-surveillance

  • 1. Gippsland School of Information Technology (GSIT) Behaviour Recognition Framework for Intelligent Visual Surveillance 06 April 2009 www.monash.edu.au
  • 2. Project Team • Mahfuzul Haque, PhD student (2 yrs, 1 m) • A/Prof. Manzur Murshed • Dr. Manoranjan Paul www.monash.edu.au 2
  • 3. Project Motivation “Behaviour Recognition Framework for Intelligent Visual Surveillance” Why “Intelligent” Surveillance? Why “Behaviour Recognition”? What type of “Behaviours”? www.monash.edu.au 3
  • 4. Surveillance Everywhere Are we really protected? www.monash.edu.au 4
  • 5. Too Many Cameras Deployment of large number of surveillance cameras in recent years London Heathrow airport has more than 5000 cameras!! www.monash.edu.au 5
  • 6. Behind the Scene: Worried Human Monitor Dependability on human monitors has increased. Reliability on surveillance system has decreased. www.monash.edu.au 6
  • 7. Project Goals  Aiding human monitors by automatic detection of specific abnormal behaviors  Decreasing dependability on human monitors  Improving reliability of surveillance systems for ensuring human security www.monash.edu.au 7
  • 8. Project Scope Group Behaviours       Mob Violence Crowding Sudden Group Formation Sudden Group Deformation Shooting Panic Driven Behaviours www.monash.edu.au 8
  • 9. Research Question How to recognize specific group behaviours from surveillance video streams in real-time? Research Area - Computer Vision - Application of Machine Learning www.monash.edu.au 9
  • 10. System Architecture Behaviour Profile Surveillance Video Stream Behaviour Recognition Framework www.monash.edu.au 10
  • 11. System Architecture Behaviour Profile Surveillance Video Stream Behaviour Recognition Framework www.monash.edu.au 11
  • 12. Behaviour Profile Surveillance Video Stream (System Input) Time 0 Unknown 20 Group Appearing 60 Group Appearing 140 320 Group Merging Group Splitting Behaviour Profile (System Output) www.monash.edu.au 12
  • 13. System Architecture Behaviour Profile Surveillance Video Stream Behaviour Recognition Framework www.monash.edu.au 13
  • 14. Behaviour Recognition Framework Framework Components • • • • Background Modelling Frame Level Feature Extraction Temporal Feature Extraction Behaviour Classification www.monash.edu.au 14
  • 15. Behaviour Recognition Framework Background Modelling Frame Level Feature Extraction Temporal Feature Extraction Behaviour Classification www.monash.edu.au 15
  • 17. Background Modelling How to extract the active regions from surveillance video stream? Background Subtraction Current frame Challenges!! = Background Moving foreground www.monash.edu.au 17
  • 18. Background Modelling σ2 P(x) µ P(x) x Sky Cloud Leaf Moving Person σ2 Road Shadow Moving Car Floor Shadow Walking People Cloud µ P(x) x P(x) Person Leaf Sky σ2 µ x x (Pixel intensity) www.monash.edu.au 18
  • 19. Background Modelling Background Model Current frame Moving foreground Frame 1 Frame N Background Models road shadow car shadow road Models are ordered by ω/σ ω1 σ12 µ1 road ω2 σ22 µ2 shadow 65% 20% ω3 σ32 µ3 car 15% www.monash.edu.au 19
  • 20. Background Modelling First Frame Test Frame Ground Truth S&G Lee Proposed (1) (2) (3) (4) (5) (1) PETS2000; (2) PETS2006-B1; (3) PETS2006-B2; (4) PETS2006-B3; and (5) PETS2006-B4. www.monash.edu.au 20
  • 21. Frame Level Feature Extraction Background Modelling Frame Level Feature Extraction Temporal Feature Extraction Behaviour Classification www.monash.edu.au 21
  • 22. Frame Level Feature Extraction • Feature Categories: – Count – Area – Density – Bounding Box – Filling Ratio – Aspect Ratio • 30 frame level features Bounding Boxes www.monash.edu.au 22
  • 23. Frame Level Feature Extraction Foreground Count Foreground Area Foreground Density • FC (Foreground Count) • TFA (Total Foreground Area) • AFA (Average Foreground Area) • VFA (Variance of Foreground Area) • MAXFA (Maximum Foreground Area) • MINFA (Minimum Foreground Area) • AFD (Average Foreground Density) • VFD (Variance of Foreground Density) Filling Ratio Bounding Box – Area Bounding Box – Width • TFR (Total Filling Ratio) • AFR (Average Filling Ratio) • VFR (Variance of Filling Ratio) • MAXFR (Maximum Filling Ratio) • MINFR (Minimum Filling Ratio) • TBBA (Total Bounding Box Area) • ABBA (Average Bounding Box Area) • VBBA (Variance of Bounding Box Area) • MAXBBA (Maximum Bounding Box Area) • MINBBA (Minimum Bounding Box Area) • ABBW (Average Bounding Box Width) • VBBW (Variance of Bounding Box Width) • MAXBBW (Maximum Bounding Box Width) • MINBBW (Minimum Bounding Box Width) Bounding Box – Height Aspect Ratio • ABBH (Average Bounding Box Height) • VBBH (Variance of Bounding Box Height) • MAXBBH (Maximum Bounding Box Height) • MINBBH (Minimum Bounding Box Height) • AAR (Average Aspect Ratio) • VAR (Variance of Aspect Ratio) • MAXAR (Maximum Aspect Ratio) • MINAR (Minimum Aspect Ratio) www.monash.edu.au 23
  • 24. Temporal Feature Extraction Background Modelling Frame Level Feature Extraction Temporal Feature Extraction Behaviour Classification www.monash.edu.au 24
  • 25. Temporal Feature Extraction • Fixed length, partially overlapped sliding window • Temporal data smoothing – polynomial curve fitting • 9 temporal features for each frame level feature • Output: 270 temporal features www.monash.edu.au 25
  • 26. Temporal Feature Extraction TFA (Total Foreground Area) Temporal Features TFA (%) • MAX • MIN • AVG • VAR • RATE • TIME(MAX) • TIME(MIN) • D = TIME(MAX) - TIME(MIN) • SLOPE ( D/2 ) Time (window = 100 frames) www.monash.edu.au 26
  • 27. Temporal Feature Extraction TFA (Total Foreground Area) Temporal Features TFA (%) MAX MIN • MAX • MIN • AVG • VAR • RATE • TIME(MAX) • TIME(MIN) • D = TIME(MAX) - TIME(MIN) • SLOPE ( D/2 ) TIME(MAX) TIME(MIN) Time (window = 100 frames) www.monash.edu.au 27
  • 29. Behaviour Classification • • • • • • • • Individual classifiers for each behaviour Supervised training Feature ranking Top 100 features from 270 features Dimension reduction (PCA) Max dimension 30 SVM classifier Output: Behaviour Profile www.monash.edu.au 29
  • 30. Behaviour Classification Experiments GROUP FORMING • Accuracy: 0.9767 • Top 3 features • TIME(MAX)-VFD • TIME(MAX)-AFD • TIME(MAX) - TIME(MIN)-VFD GROUP SPLITTING AND SPREADING • Accuracy: 0.8488 • Top 3 features • TIME(MAX)-VFD • TIME(MIN)-ABBA • TIME(MIN)-AFA BLOCKED EXIT • Accuracy: 0.9651 • Top 3 features • TIME(MIN)-TFA • MIN-MINAR • TIME(MAX)-TFA www.monash.edu.au 30
  • 31. Summary: Framework Components Background Modelling Frame Level Feature Extraction Temporal Feature Extraction Behaviour Classification • Multiple Background Models • Gaussian Mixture Models (GMM) • Unsupervised • Output: Foreground Region/Mask • Feature Categories: • Count • Area • Density • Bounding Box • Filling Ratio • Aspect Ratio • Output: 30 Frame Level Features • Fixed Length, Partially Overlapped Sliding Window • Temporal Data Smoothing – Polynomial Curve Fitting • 9 Temporal Features for Each Frame Level Features • Output: 270 Temporal Features • Individual Classifiers for Each Behaviour • Each Classifier is Trained Using Supervised Learning • Feature Ranking • Top 100 Features • Dimension Reduction (PCA) • Max Dimension 30 • SVM classifier • Output: Behaviour Profile www.monash.edu.au 31
  • 32. Summary: Framework Components Background Modelling Frame Level Feature Extraction Temporal Feature Extraction Behaviour Classification • Multiple Background Models • Gaussian Mixture Models (GMM) • Unsupervised • Output: Foreground Region/Mask • Feature Categories: • Count • Area • Density • Bounding Box • Filling Ratio • Aspect Ratio • Output: 30 Frame Level Features • Fixed Length, Partially Overlapped Sliding Window • Temporal Data Smoothing – Polynomial Curve Fitting • 9 Temporal Features for Each Frame Level Features • Output: 270 Temporal Features • Individual Classifiers for Each Behaviour • Each Classifier is Trained Using Supervised Learning • Feature Ranking • Top 100 Features • Dimension Reduction (PCA) • Max Dimension 30 • SVM classifier • Output: Behaviour Profile www.monash.edu.au 32
  • 33. Research Challenges • No tracking/trajectory • Simple behaviours – Group appear/disappear – Group merge/split • Panic driven behaviours – Fire/Blocked exit – Fighting/Shooting • Context variation – Speed – Direction – Object Resolution www.monash.edu.au 33
  • 35. Publications 1. Mahfuzul Haque, Manzur Murshed, and Manoranjan Paul, Improved Gaussian Mixtures for Robust Object Detection by Adaptive MultiBackground Generation, International Conference on Pattern Recognition (ICPR), Tampa, Florida, USA, 2008. (CORE A) 2. Mahfuzul Haque, Manzur Murshed, and Manoranjan Paul, A Hybrid Object Detection Technique from Dynamic Background Using Gaussian Mixture Models, IEEE International Workshop on Multimedia Signal Processing (MMSP), Cairns, Australia, 2008. (CORE A) 3. Mahfuzul Haque, Manzur Murshed, and Manoranjan Paul, On Stable Dynamic Background Generation Technique using Gaussian Mixture Models for Robust Object Detection, IEEE International Conference On Advanced Video and Signal Based Surveillance (AVSS), Santa Fe, New Mexico, USA, 2008. (CORE A) CORE - COmputing Research and Education Association www.monash.edu.au 35
  • 37. Acknowledgments URLs of the images used in this presentation • • • • • • • • • • • • • • http://www.fotosearch.com/DGV464/766029/ http://www.cyprus-trader.com/images/alert.gif http://security.polito.it/~lioy/img/einstein8ci.jpg http://www.dtsc.ca.gov/PollutionPrevention/images/question.jpg http://www.unmikonline.org/civpol/photos/thematic/violence/streetvio2.jpg http://www.airports-worldwide.com/img/uk/heathrow00.jpg http://www.highprogrammer.com/alan/gaming/cons/trips/genconindy2003/exhibithall-crowd-2.jpg http://www.bhopal.org/fcunited/archives/fcu-crowd.jpg http://img.dailymail.co.uk/i/pix/2006/08/passaPA_450x300.jpg http://www.defenestrator.org/drp/files/surveillance-cameras-400.jpg http://www.cityofsound.com/photos/centre_poin/crowd.jpg http://www.hindu.com/2007/08/31/images/2007083156401501.jpg http://paulaoffutt.com/pics/images/crowd-surfing.jpg http://msnbcmedia1.msn.com/j/msnbc/Components/Photos/070225/070225_surv eillance_hmed.hmedium.jpg www.monash.edu.au 37