SlideShare a Scribd company logo
1 of 30
Download to read offline
Information Technology

Abnormal Event Detection in
Unseen Scenarios
Mahfuzul Haque and Manzur Murshed
Outline
 Event Detection for Public Safety
 Challenges
 Proposed Approach

 Experiments
 Summary

 Q&A

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

2
Event Detection for Public Safety
 Mob Violence
 Crowding
 Sudden Group Formation
 Sudden Group Deformation
 Shooting

 Panic Driven Behaviours

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

3
Event Detection

time





Specific event (e.g., run) detection vs. abnormality detection
An event persists for a certain duration of time
The duration is variable
The characteristics of the same event is
 variable in the same environment
 variable from one scene to other
Abnormal Event Detection in Unseen Scenarios

December 30, 2013

4
Challenges

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

5
Challenges
Abnormal Event Detection

Supervised

Unsupervised

Manual Labelling,
Prior assumption of well
define event classes

No Event Model
Clustering of observed patterns,
Database of spatiotemporal patches

Semi-supervised
Normal event modelling:
manual labelling,
Abnormal event modelling:
unsupervised adaptation

Explicit Event Model
More Recent Approach
Mixture of Dynamic Bayesian Networks

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

6
Challenges

Build Event Model Once

Operate Everywhere
Abnormal Event Detection in Unseen Scenarios

December 30, 2013

7
Proposed Approach
Build
•
•
•
•

Targeted Events
Extensive Feature Extraction
Feature Selection/Ranking
Supervised

Operation
•
•
•
•
•

No additional training
No parameter tuning
Selected feature extraction
No feature ranking
Real-time detection

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

8
Proposed Approach
f1
f2

Event

f3

Model

.

time
Frame-level Features

fn
Temporal Features

Classifier

 Event detection as temporal data classification problem
 A distinct set of temporal features can characterise an event
 Independent frame-level features extracted using blob statistical
analysis; no object / position specific information, no spatial
association
 Frame-level features are transformed into temporal features
considering speed and temporal order
Abnormal Event Detection in Unseen Scenarios

December 30, 2013

9
Proposed Approach
Processes
Foreground
Detector

Frame-level
Feature Extractor

Temporal
Feature Extractor

Event
Models

Model Training (offline)
Frame-level
Feature Extraction
(30 features)

Background
Subtraction
Labelled frames

Temporal
Feature Extraction
(270 features)

Feature Ranking
and Selection

Event Model
Training

Foreground blobs

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

10
Blob-Statistical Analysis
Frame-level features










Blob Area (BA)
Filling Ratio (FR)
Aspect Ratio (AR)
Bounding Box Area (BBA)
Bounding box Width (BBW)
Bounding box Height (BBH)
Blob Count (BC)
Blob Distance (BD)

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

11
Blob Statistical Analysis
Blob Count (BC), Blob Area (BA)

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

12
Blob Statistical Analysis
Blob Distance (BD)

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

13
Blob Statistical Analysis
Aspect Ratio (AR)

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

14
Feature Extraction
Temporal features
2
1

4
3

6
5

Frame #

 Overlapping sliding window
 Temporal order
 Speed of variation

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

15
Feature Extraction
Top five features for four different events

Feature ranking using absolute value criteria of two sample t-test, based on
pooled variance estimate.
Abnormal Event Detection in Unseen Scenarios

December 30, 2013

16
Proposed Approach
Model Training (offline)
Frame-level
Feature Extraction
(30 features)

Background
Subtraction
Labelled frames

Temporal
Feature Extraction
(270 features)

Feature Ranking
and Selection

Event Model
Training

Foreground blobs

Event Detection in the Operating Environment
Selective
Frame-level
Feature Extraction

Background
Subtraction
Incoming frames

Selective
Temporal
Feature Extraction

Trained
Event Models

Detection
Results

Foreground blobs

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

17
Summary and Discussion
Motion based approaches
 Key points detection
 Point matching in successive frames
 Flow vectors: position, direction, speed

Tracking based approaches
 Object detection
 Object matching in successive frames
 Trajectories: object paths

Common characteristics
 Inter-frame association
 Context specific information
 Event models are not generic
Hu et al. (ICPR 2008)

Xiang et al. (IJCV 2006)

Proposed approach

 No Inter-frame association
 Foreground blob detection
 Independent frame-level features =>
 Global frame-level descriptor based on
temporal features considering speed
blob statistical analysis, independent
and temporal order
of scene characteristics
Abnormal Event Detection in Unseen Scenarios

December 30, 2013

18
Experiments

Model Training
•
•
•
•
•
•
•

Four different events: meet, split, runaway, and fight
CAVIAR dataset with labelled frames
80% of the test frames for model training
100 iterations of 10-fold cross validation
Remaining 20% of the test frames for testing
SVM classifier as event models
Separate model for each event

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

19
Experiments
Event Models

Unseen Scenarios in
Known Context

Unseen Scenarios in
Unknown Context

Greenfield

Outdoor

Abnormal Event Detection in Unseen Scenarios

Corridor

December 30, 2013

20
Experiments

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

21
Experiments

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

22
Experiments
• Abnormal event detection in unseen scenarios in
unknown context
• University of Minnesota crowd dataset (UMN dataset)
• The Runaway event model
• No additional training or tuning
• Three different sites

Greenfield

Outdoor

Corridor

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

23
Experiments
Abnormal Event Detection (UMN-9)

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

24
Experiments
Abnormal Event Detection (UMN-10)

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

25
Experiments
Abnormal Event Detection (UMN-01)

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

26
Experiments
Abnormal Event Detection (UMN-07)

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

27
Experiments

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

28
Experiments

Method

AUC

Proposed Method

0.89

Pure Optical Flow [1]

0.84

[1] R. Mehran, A. Oyama, and M. Shah, “Abnormal crowd behavior detection using social force model,” in Proc. IEEE
Conference on Computer Vision and Pattern Recognition CVPR 2009, Event Detection2009, pp.Scenarios
Abnormal 20–25 June in Unseen 935–942. December 30, 2013

29
Q&A

Abnormal Event Detection in Unseen Scenarios

December 30, 2013

30

More Related Content

More from Mahfuzul Haque

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

More from Mahfuzul Haque (16)

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

Recently uploaded

Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...
Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...
Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...
panagenda
 
Breaking Down the Flutterwave Scandal What You Need to Know.pdf
Breaking Down the Flutterwave Scandal What You Need to Know.pdfBreaking Down the Flutterwave Scandal What You Need to Know.pdf
Breaking Down the Flutterwave Scandal What You Need to Know.pdf
UK Journal
 

Recently uploaded (20)

Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdfLinux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
 
Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...
Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...
Choosing the Right FDO Deployment Model for Your Application _ Geoffrey at In...
 
WebAssembly is Key to Better LLM Performance
WebAssembly is Key to Better LLM PerformanceWebAssembly is Key to Better LLM Performance
WebAssembly is Key to Better LLM Performance
 
Extensible Python: Robustness through Addition - PyCon 2024
Extensible Python: Robustness through Addition - PyCon 2024Extensible Python: Robustness through Addition - PyCon 2024
Extensible Python: Robustness through Addition - PyCon 2024
 
Long journey of Ruby Standard library at RubyKaigi 2024
Long journey of Ruby Standard library at RubyKaigi 2024Long journey of Ruby Standard library at RubyKaigi 2024
Long journey of Ruby Standard library at RubyKaigi 2024
 
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
 
Where to Learn More About FDO _ Richard at FIDO Alliance.pdf
Where to Learn More About FDO _ Richard at FIDO Alliance.pdfWhere to Learn More About FDO _ Richard at FIDO Alliance.pdf
Where to Learn More About FDO _ Richard at FIDO Alliance.pdf
 
Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...
Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...
Easier, Faster, and More Powerful – Alles Neu macht der Mai -Wir durchleuchte...
 
Simplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdf
Simplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdfSimplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdf
Simplified FDO Manufacturing Flow with TPMs _ Liam at Infineon.pdf
 
Enterprise Knowledge Graphs - Data Summit 2024
Enterprise Knowledge Graphs - Data Summit 2024Enterprise Knowledge Graphs - Data Summit 2024
Enterprise Knowledge Graphs - Data Summit 2024
 
Your enemies use GenAI too - staying ahead of fraud with Neo4j
Your enemies use GenAI too - staying ahead of fraud with Neo4jYour enemies use GenAI too - staying ahead of fraud with Neo4j
Your enemies use GenAI too - staying ahead of fraud with Neo4j
 
WSO2CONMay2024OpenSourceConferenceDebrief.pptx
WSO2CONMay2024OpenSourceConferenceDebrief.pptxWSO2CONMay2024OpenSourceConferenceDebrief.pptx
WSO2CONMay2024OpenSourceConferenceDebrief.pptx
 
Integrating Telephony Systems with Salesforce: Insights and Considerations, B...
Integrating Telephony Systems with Salesforce: Insights and Considerations, B...Integrating Telephony Systems with Salesforce: Insights and Considerations, B...
Integrating Telephony Systems with Salesforce: Insights and Considerations, B...
 
Microsoft CSP Briefing Pre-Engagement - Questionnaire
Microsoft CSP Briefing Pre-Engagement - QuestionnaireMicrosoft CSP Briefing Pre-Engagement - Questionnaire
Microsoft CSP Briefing Pre-Engagement - Questionnaire
 
Breaking Down the Flutterwave Scandal What You Need to Know.pdf
Breaking Down the Flutterwave Scandal What You Need to Know.pdfBreaking Down the Flutterwave Scandal What You Need to Know.pdf
Breaking Down the Flutterwave Scandal What You Need to Know.pdf
 
Designing for Hardware Accessibility at Comcast
Designing for Hardware Accessibility at ComcastDesigning for Hardware Accessibility at Comcast
Designing for Hardware Accessibility at Comcast
 
IESVE for Early Stage Design and Planning
IESVE for Early Stage Design and PlanningIESVE for Early Stage Design and Planning
IESVE for Early Stage Design and Planning
 
The Metaverse: Are We There Yet?
The  Metaverse:    Are   We  There  Yet?The  Metaverse:    Are   We  There  Yet?
The Metaverse: Are We There Yet?
 
Powerful Start- the Key to Project Success, Barbara Laskowska
Powerful Start- the Key to Project Success, Barbara LaskowskaPowerful Start- the Key to Project Success, Barbara Laskowska
Powerful Start- the Key to Project Success, Barbara Laskowska
 
BT & Neo4j _ How Knowledge Graphs help BT deliver Digital Transformation.pptx
BT & Neo4j _ How Knowledge Graphs help BT deliver Digital Transformation.pptxBT & Neo4j _ How Knowledge Graphs help BT deliver Digital Transformation.pptx
BT & Neo4j _ How Knowledge Graphs help BT deliver Digital Transformation.pptx
 

Talk 2012-icmew-event

  • 1. Information Technology Abnormal Event Detection in Unseen Scenarios Mahfuzul Haque and Manzur Murshed
  • 2. Outline  Event Detection for Public Safety  Challenges  Proposed Approach  Experiments  Summary  Q&A Abnormal Event Detection in Unseen Scenarios December 30, 2013 2
  • 3. Event Detection for Public Safety  Mob Violence  Crowding  Sudden Group Formation  Sudden Group Deformation  Shooting  Panic Driven Behaviours Abnormal Event Detection in Unseen Scenarios December 30, 2013 3
  • 4. Event Detection time     Specific event (e.g., run) detection vs. abnormality detection An event persists for a certain duration of time The duration is variable The characteristics of the same event is  variable in the same environment  variable from one scene to other Abnormal Event Detection in Unseen Scenarios December 30, 2013 4
  • 5. Challenges Abnormal Event Detection in Unseen Scenarios December 30, 2013 5
  • 6. Challenges Abnormal Event Detection Supervised Unsupervised Manual Labelling, Prior assumption of well define event classes No Event Model Clustering of observed patterns, Database of spatiotemporal patches Semi-supervised Normal event modelling: manual labelling, Abnormal event modelling: unsupervised adaptation Explicit Event Model More Recent Approach Mixture of Dynamic Bayesian Networks Abnormal Event Detection in Unseen Scenarios December 30, 2013 6
  • 7. Challenges Build Event Model Once Operate Everywhere Abnormal Event Detection in Unseen Scenarios December 30, 2013 7
  • 8. Proposed Approach Build • • • • Targeted Events Extensive Feature Extraction Feature Selection/Ranking Supervised Operation • • • • • No additional training No parameter tuning Selected feature extraction No feature ranking Real-time detection Abnormal Event Detection in Unseen Scenarios December 30, 2013 8
  • 9. Proposed Approach f1 f2 Event f3 Model . time Frame-level Features fn Temporal Features Classifier  Event detection as temporal data classification problem  A distinct set of temporal features can characterise an event  Independent frame-level features extracted using blob statistical analysis; no object / position specific information, no spatial association  Frame-level features are transformed into temporal features considering speed and temporal order Abnormal Event Detection in Unseen Scenarios December 30, 2013 9
  • 10. Proposed Approach Processes Foreground Detector Frame-level Feature Extractor Temporal Feature Extractor Event Models Model Training (offline) Frame-level Feature Extraction (30 features) Background Subtraction Labelled frames Temporal Feature Extraction (270 features) Feature Ranking and Selection Event Model Training Foreground blobs Abnormal Event Detection in Unseen Scenarios December 30, 2013 10
  • 11. Blob-Statistical Analysis Frame-level features         Blob Area (BA) Filling Ratio (FR) Aspect Ratio (AR) Bounding Box Area (BBA) Bounding box Width (BBW) Bounding box Height (BBH) Blob Count (BC) Blob Distance (BD) Abnormal Event Detection in Unseen Scenarios December 30, 2013 11
  • 12. Blob Statistical Analysis Blob Count (BC), Blob Area (BA) Abnormal Event Detection in Unseen Scenarios December 30, 2013 12
  • 13. Blob Statistical Analysis Blob Distance (BD) Abnormal Event Detection in Unseen Scenarios December 30, 2013 13
  • 14. Blob Statistical Analysis Aspect Ratio (AR) Abnormal Event Detection in Unseen Scenarios December 30, 2013 14
  • 15. Feature Extraction Temporal features 2 1 4 3 6 5 Frame #  Overlapping sliding window  Temporal order  Speed of variation Abnormal Event Detection in Unseen Scenarios December 30, 2013 15
  • 16. Feature Extraction Top five features for four different events Feature ranking using absolute value criteria of two sample t-test, based on pooled variance estimate. Abnormal Event Detection in Unseen Scenarios December 30, 2013 16
  • 17. Proposed Approach Model Training (offline) Frame-level Feature Extraction (30 features) Background Subtraction Labelled frames Temporal Feature Extraction (270 features) Feature Ranking and Selection Event Model Training Foreground blobs Event Detection in the Operating Environment Selective Frame-level Feature Extraction Background Subtraction Incoming frames Selective Temporal Feature Extraction Trained Event Models Detection Results Foreground blobs Abnormal Event Detection in Unseen Scenarios December 30, 2013 17
  • 18. Summary and Discussion Motion based approaches  Key points detection  Point matching in successive frames  Flow vectors: position, direction, speed Tracking based approaches  Object detection  Object matching in successive frames  Trajectories: object paths Common characteristics  Inter-frame association  Context specific information  Event models are not generic Hu et al. (ICPR 2008) Xiang et al. (IJCV 2006) Proposed approach  No Inter-frame association  Foreground blob detection  Independent frame-level features =>  Global frame-level descriptor based on temporal features considering speed blob statistical analysis, independent and temporal order of scene characteristics Abnormal Event Detection in Unseen Scenarios December 30, 2013 18
  • 19. Experiments Model Training • • • • • • • Four different events: meet, split, runaway, and fight CAVIAR dataset with labelled frames 80% of the test frames for model training 100 iterations of 10-fold cross validation Remaining 20% of the test frames for testing SVM classifier as event models Separate model for each event Abnormal Event Detection in Unseen Scenarios December 30, 2013 19
  • 20. Experiments Event Models Unseen Scenarios in Known Context Unseen Scenarios in Unknown Context Greenfield Outdoor Abnormal Event Detection in Unseen Scenarios Corridor December 30, 2013 20
  • 21. Experiments Abnormal Event Detection in Unseen Scenarios December 30, 2013 21
  • 22. Experiments Abnormal Event Detection in Unseen Scenarios December 30, 2013 22
  • 23. Experiments • Abnormal event detection in unseen scenarios in unknown context • University of Minnesota crowd dataset (UMN dataset) • The Runaway event model • No additional training or tuning • Three different sites Greenfield Outdoor Corridor Abnormal Event Detection in Unseen Scenarios December 30, 2013 23
  • 24. Experiments Abnormal Event Detection (UMN-9) Abnormal Event Detection in Unseen Scenarios December 30, 2013 24
  • 25. Experiments Abnormal Event Detection (UMN-10) Abnormal Event Detection in Unseen Scenarios December 30, 2013 25
  • 26. Experiments Abnormal Event Detection (UMN-01) Abnormal Event Detection in Unseen Scenarios December 30, 2013 26
  • 27. Experiments Abnormal Event Detection (UMN-07) Abnormal Event Detection in Unseen Scenarios December 30, 2013 27
  • 28. Experiments Abnormal Event Detection in Unseen Scenarios December 30, 2013 28
  • 29. Experiments Method AUC Proposed Method 0.89 Pure Optical Flow [1] 0.84 [1] R. Mehran, A. Oyama, and M. Shah, “Abnormal crowd behavior detection using social force model,” in Proc. IEEE Conference on Computer Vision and Pattern Recognition CVPR 2009, Event Detection2009, pp.Scenarios Abnormal 20–25 June in Unseen 935–942. December 30, 2013 29
  • 30. Q&A Abnormal Event Detection in Unseen Scenarios December 30, 2013 30