SlideShare ist ein Scribd-Unternehmen logo
1 von 12
Yusuf Ziya Uzun
Artificial Neural Networks – CMP5133
 Resample video to desired size
 Divide video to Images
 Capture OF (Optical Flow) Vectors
 DBSCAN clustering
 Find OF Vector orientation
 Colorize clusters by using vector orientations
 The ground truth does not exist: The desired
results always depend on the user
requirements and specifications.
 Even for a fixed image, there may be more
than one "best" segmentation because the
criteria defining the quality of a
segmentation are application dependent.
-Pierre Soille
 Motion: displacement, direction, velocity,
acceleration, time and speed
 Optical Flow: distribution of the apparent
velocities of objects in an image
Zoom out Zoom in Pan right to left
 Two Main Category: Sparse and Dense
 Horn and Schunck
 Kanade-Lucas-Tomasi(KLT)
 Gunnar - Farneback
 Separate moving objects from background by
using motion vectors(optical flow)
 Just split image N pieces.
 Problems:
 Aperture
 Barber-pole (Motion vs Optical)
 Closer Objects Have Bigger Velocity?
 Stereo Vision
 Density-based spatial clustering of
applications with noise (DBSCAN)
 Given a set of points and radius:
 Groups close points
 Alone points become outliers
 C# and EmguCV
 Resampling video with ffmpeg manually
 Ratio: same in video
 Size: 640 px width
 Divide video and capture frames (x – 5) and x
to compare
 OF Vectors:
 Gunnar – Farneback Dense OF Vectors
 Gaussian Box Filter
 A global threshold to remove noise
 DBSCAN:
 Globally defined epsilon and # of points
 Computing clusters of OF vectors
 OF vector orientation
 Coloring clusters by looking OF vector
orientations
 Many Global Variables
 DBSCAN and OF combination useful
 Experimental
 Variables Domain Dependent
 Not good to use everywhere
 Can combine with Supervised Learning
An Application of Video Segmentation Using Optical Flows

Weitere ähnliche Inhalte

Was ist angesagt?

Temporal Frequency Probing for 5D Transient Analysis of Global Light Transport
Temporal Frequency Probing for 5D Transient Analysis of Global Light TransportTemporal Frequency Probing for 5D Transient Analysis of Global Light Transport
Temporal Frequency Probing for 5D Transient Analysis of Global Light Transport
Matthew O'Toole
 
Ibica2014(p15)image fusion based on broveywavelet
Ibica2014(p15)image fusion based on broveywaveletIbica2014(p15)image fusion based on broveywavelet
Ibica2014(p15)image fusion based on broveywavelet
Aboul Ella Hassanien
 
Photography as foundation of cinematography
Photography as foundation of cinematographyPhotography as foundation of cinematography
Photography as foundation of cinematography
Ivy Autor
 

Was ist angesagt? (20)

Light Field Technology
Light Field TechnologyLight Field Technology
Light Field Technology
 
Introduction to Light Fields
Introduction to Light FieldsIntroduction to Light Fields
Introduction to Light Fields
 
Light Field Photography Introduction
Light Field Photography IntroductionLight Field Photography Introduction
Light Field Photography Introduction
 
IJRET-V1I1P2 -A Survey Paper On Single Image and Video Dehazing Methods
IJRET-V1I1P2 -A Survey Paper On Single Image and Video Dehazing MethodsIJRET-V1I1P2 -A Survey Paper On Single Image and Video Dehazing Methods
IJRET-V1I1P2 -A Survey Paper On Single Image and Video Dehazing Methods
 
SIGGRAPH 2012 Computational Plenoptic Imaging Course - 4 Light Fields
SIGGRAPH 2012 Computational Plenoptic Imaging Course - 4 Light FieldsSIGGRAPH 2012 Computational Plenoptic Imaging Course - 4 Light Fields
SIGGRAPH 2012 Computational Plenoptic Imaging Course - 4 Light Fields
 
Dense Image Matching - Challenges and Potentials (Keynote 3D-ARCH 2015)
Dense Image Matching - Challenges and Potentials (Keynote 3D-ARCH 2015) Dense Image Matching - Challenges and Potentials (Keynote 3D-ARCH 2015)
Dense Image Matching - Challenges and Potentials (Keynote 3D-ARCH 2015)
 
A Review on Airlight Estimation Haze Removal Algorithms
A Review on Airlight Estimation Haze Removal AlgorithmsA Review on Airlight Estimation Haze Removal Algorithms
A Review on Airlight Estimation Haze Removal Algorithms
 
HDR in Cinema: Achievable Contrast
HDR in Cinema: Achievable Contrast HDR in Cinema: Achievable Contrast
HDR in Cinema: Achievable Contrast
 
Voxel based global-illumination
Voxel based global-illuminationVoxel based global-illumination
Voxel based global-illumination
 
Temporal Frequency Probing for 5D Transient Analysis of Global Light Transport
Temporal Frequency Probing for 5D Transient Analysis of Global Light TransportTemporal Frequency Probing for 5D Transient Analysis of Global Light Transport
Temporal Frequency Probing for 5D Transient Analysis of Global Light Transport
 
3D Shape and Indirect Appearance by Structured Light Transport
3D Shape and Indirect Appearance by Structured Light Transport3D Shape and Indirect Appearance by Structured Light Transport
3D Shape and Indirect Appearance by Structured Light Transport
 
Ibica2014(p15)image fusion based on broveywavelet
Ibica2014(p15)image fusion based on broveywaveletIbica2014(p15)image fusion based on broveywavelet
Ibica2014(p15)image fusion based on broveywavelet
 
Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...
Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...
Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...
 
SIGGRAPH 2014 Course on Computational Cameras and Displays (part 2)
SIGGRAPH 2014 Course on Computational Cameras and Displays (part 2)SIGGRAPH 2014 Course on Computational Cameras and Displays (part 2)
SIGGRAPH 2014 Course on Computational Cameras and Displays (part 2)
 
HDR
HDRHDR
HDR
 
Coded Photography - Ramesh Raskar
Coded Photography - Ramesh RaskarCoded Photography - Ramesh Raskar
Coded Photography - Ramesh Raskar
 
Primal-Dual Coding to Probe Light Transport
Primal-Dual Coding to Probe Light TransportPrimal-Dual Coding to Probe Light Transport
Primal-Dual Coding to Probe Light Transport
 
Compressed Sensing - Achuta Kadambi
Compressed Sensing - Achuta KadambiCompressed Sensing - Achuta Kadambi
Compressed Sensing - Achuta Kadambi
 
Photography as foundation of cinematography
Photography as foundation of cinematographyPhotography as foundation of cinematography
Photography as foundation of cinematography
 
mihara_iccp16_presentation
mihara_iccp16_presentationmihara_iccp16_presentation
mihara_iccp16_presentation
 

Ähnlich wie An Application of Video Segmentation Using Optical Flows

Troubleshooting, Designing & Installing Digital & Analog Closed Circuit TV Sy...
Troubleshooting, Designing & Installing Digital & Analog Closed Circuit TV Sy...Troubleshooting, Designing & Installing Digital & Analog Closed Circuit TV Sy...
Troubleshooting, Designing & Installing Digital & Analog Closed Circuit TV Sy...
Living Online
 
ESPI - Project Report
ESPI - Project Report ESPI - Project Report
ESPI - Project Report
Akash Marakani
 
Alternatives to Point-Scan Confocal Microscopy
Alternatives to Point-Scan Confocal MicroscopyAlternatives to Point-Scan Confocal Microscopy
Alternatives to Point-Scan Confocal Microscopy
mchelen
 
Human Action Recognition Based on Spacio-temporal features
Human Action Recognition Based on Spacio-temporal featuresHuman Action Recognition Based on Spacio-temporal features
Human Action Recognition Based on Spacio-temporal features
nikhilus85
 
Iaetsd deblurring of noisy or blurred
Iaetsd deblurring of noisy or blurredIaetsd deblurring of noisy or blurred
Iaetsd deblurring of noisy or blurred
Iaetsd Iaetsd
 

Ähnlich wie An Application of Video Segmentation Using Optical Flows (20)

Troubleshooting, Designing & Installing Digital & Analog Closed Circuit TV Sy...
Troubleshooting, Designing & Installing Digital & Analog Closed Circuit TV Sy...Troubleshooting, Designing & Installing Digital & Analog Closed Circuit TV Sy...
Troubleshooting, Designing & Installing Digital & Analog Closed Circuit TV Sy...
 
MOTION FLOW
MOTION FLOWMOTION FLOW
MOTION FLOW
 
Back projection geometry in cbct
Back projection geometry in cbctBack projection geometry in cbct
Back projection geometry in cbct
 
Keynote at Tracking Workshop during ISMAR 2014
Keynote at Tracking Workshop during ISMAR 2014Keynote at Tracking Workshop during ISMAR 2014
Keynote at Tracking Workshop during ISMAR 2014
 
Dsc
DscDsc
Dsc
 
Dsc
DscDsc
Dsc
 
Dsc
DscDsc
Dsc
 
Dsc
DscDsc
Dsc
 
presentation.ppt
presentation.pptpresentation.ppt
presentation.ppt
 
New microsoft power point presentation
New microsoft power point presentationNew microsoft power point presentation
New microsoft power point presentation
 
IJSRDV3I40293
IJSRDV3I40293IJSRDV3I40293
IJSRDV3I40293
 
Action_recognition-topic.pptx
Action_recognition-topic.pptxAction_recognition-topic.pptx
Action_recognition-topic.pptx
 
ESPI - Project Report
ESPI - Project Report ESPI - Project Report
ESPI - Project Report
 
Alternatives to Point-Scan Confocal Microscopy
Alternatives to Point-Scan Confocal MicroscopyAlternatives to Point-Scan Confocal Microscopy
Alternatives to Point-Scan Confocal Microscopy
 
Deep Learning for Structure-from-Motion (SfM)
Deep Learning for Structure-from-Motion (SfM)Deep Learning for Structure-from-Motion (SfM)
Deep Learning for Structure-from-Motion (SfM)
 
Csit3916
Csit3916Csit3916
Csit3916
 
VIDEO SEGMENTATION FOR MOVING OBJECT DETECTION USING LOCAL CHANGE & ENTROPY B...
VIDEO SEGMENTATION FOR MOVING OBJECT DETECTION USING LOCAL CHANGE & ENTROPY B...VIDEO SEGMENTATION FOR MOVING OBJECT DETECTION USING LOCAL CHANGE & ENTROPY B...
VIDEO SEGMENTATION FOR MOVING OBJECT DETECTION USING LOCAL CHANGE & ENTROPY B...
 
CV_Chap 6 Motion Representation
CV_Chap 6 Motion RepresentationCV_Chap 6 Motion Representation
CV_Chap 6 Motion Representation
 
Human Action Recognition Based on Spacio-temporal features
Human Action Recognition Based on Spacio-temporal featuresHuman Action Recognition Based on Spacio-temporal features
Human Action Recognition Based on Spacio-temporal features
 
Iaetsd deblurring of noisy or blurred
Iaetsd deblurring of noisy or blurredIaetsd deblurring of noisy or blurred
Iaetsd deblurring of noisy or blurred
 

An Application of Video Segmentation Using Optical Flows

  • 1. Yusuf Ziya Uzun Artificial Neural Networks – CMP5133
  • 2.  Resample video to desired size  Divide video to Images  Capture OF (Optical Flow) Vectors  DBSCAN clustering  Find OF Vector orientation  Colorize clusters by using vector orientations
  • 3.  The ground truth does not exist: The desired results always depend on the user requirements and specifications.  Even for a fixed image, there may be more than one "best" segmentation because the criteria defining the quality of a segmentation are application dependent. -Pierre Soille
  • 4.  Motion: displacement, direction, velocity, acceleration, time and speed  Optical Flow: distribution of the apparent velocities of objects in an image Zoom out Zoom in Pan right to left
  • 5.  Two Main Category: Sparse and Dense  Horn and Schunck  Kanade-Lucas-Tomasi(KLT)  Gunnar - Farneback
  • 6.  Separate moving objects from background by using motion vectors(optical flow)  Just split image N pieces.  Problems:  Aperture  Barber-pole (Motion vs Optical)  Closer Objects Have Bigger Velocity?  Stereo Vision
  • 7.  Density-based spatial clustering of applications with noise (DBSCAN)  Given a set of points and radius:  Groups close points  Alone points become outliers
  • 8.  C# and EmguCV  Resampling video with ffmpeg manually  Ratio: same in video  Size: 640 px width  Divide video and capture frames (x – 5) and x to compare  OF Vectors:  Gunnar – Farneback Dense OF Vectors  Gaussian Box Filter  A global threshold to remove noise
  • 9.  DBSCAN:  Globally defined epsilon and # of points  Computing clusters of OF vectors  OF vector orientation  Coloring clusters by looking OF vector orientations
  • 10.
  • 11.  Many Global Variables  DBSCAN and OF combination useful  Experimental  Variables Domain Dependent  Not good to use everywhere  Can combine with Supervised Learning