SlideShare ist ein Scribd-Unternehmen logo
1 von 37
OPTICAL FLOW
Team 13
Phạm Quang Anh
Nguyễn Đức Quang
Nguyễn Quang Anh
Đỗ Văn Anh
1
Table of contents
1. Problem Definition
2. Types of Optical Flow
3. Sparse Optical Flow – Lucas-Kanade Algorithm
4. Dense Optical Flow – Horn-Schunck Algorithm
5. Conclusion
2
Problem Definition
3
1. Problem Definition
• Optical Flow: Optical flow is the motion of objects between consecutive
frames of sequence, caused by the relative movement between the object
and camera.
• Problem: Given two consecutive image frames, estimate the motion of each
pixel.
4
1. Problem Definition
5
1. Problem Definition
6
2. Types of Optical Flow
• Sparse Optical Flow: Sparse optical flow gives the flow vectors of some
"interesting features" (say few pixels depicting the edges or corners of an
object) within the frame.
7
2. Types of Optical Flow
• Dense Optical Flow: Dense optical flow attempts to compute the optical flow
vector for every pixel of each frame.
• While such computation may be slower, it gives a more accurate result and a
denser result suitable for applications such as learning structure from
motion and video segmentation.
8
Sparse Optical Flow
-
Lucas-Kanade Algorithm
9
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Lucas and Kanade proposed an effective technique to estimate the motion of
interesting features by comparing two consecutive frames.
• The Lucas-Kanade method works under the brightness constancy assumption
and small motion assumption.
10
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Shi-Tomashi Corner Detection: For the implementation of sparse optical flow,
we only track the motion of a feature set of pixels. Features in images are
points of interest which present rich image content information.
11
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Shi-Tomashi Corner Detection:
12
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Shi-Tomashi Corner Detection:
13
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Shi-Tomashi Corner Detection:
14
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Brightness constancy assumption: Brightness of the point will remain the
same.
I(x(t), y(t), t) = C - constant
15
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Small motion assumption:
16
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Small motion assumption:
17
3. Sparse Optical Flow – Lucas-Kanade Algorithm
18
3. Sparse Optical Flow – Lucas-Kanade Algorithm
19
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Aperture problem:
20
3. Sparse Optical Flow – Lucas-Kanade Algorithm
21
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Smoothness constancy assumption: A frame portrays a “natural” scene with
textured objects exhibiting shades of gray that change smoothly.
22
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Smoothness constancy assumption:
23
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Smoothness constancy assumption:
24
3. Sparse Optical Flow – Lucas-Kanade Algorithm
• Smoothness constancy assumption:
25
Dense Optical Flow
-
Horn-Schuck Algorithm
26
4. Dense Optical Flow – Horn-Schunck Algorithm
• The Horn–Schunck method of estimating optical flow is a global method
which introduces a global constraint of smoothness to solve the aperture
problem.
27
4. Dense Optical Flow – Horn-Schunck Algorithm
28
4. Dense Optical Flow – Horn-Schunck Algorithm
29
4. Dense Optical Flow – Horn-Schunck Algorithm
30
4. Dense Optical Flow – Horn-Schunck Algorithm
31
LUCAS-KANADE WITH PYRAMID
32
5. Lucas Kanade with Pyramid
33
In regular optical flow method, we assume the following:
• Brightness constancy
• Small motion
• Spatial conherence
If the object were to move a larger distance
→The traditional optical flow method would work bad
5. Lucas Kanade with Pyramid
34
• Pyramid is built by using multiple copies
of the same image.
• Each level in the pyramid is 1/4th of the
size of the previous level
• The lowest level is of the highest
resolution
• The highest level is of the lowest
resolution
5. Lucas Kanade with Pyramid
35
• Pyramid is built by using multiple copies of the same image.
• Each level in the pyramid is 1/4th of the size of the previous level
• The lowest level is of the highest resolution
• The highest level is of the lowest resolution
• To Downsample: Using Gausian pyramid
• To Upsamgple: Using Laplacian pyramid
5. Lucas Kanade with Pyramid
36
Lucas-Kanade with Pyramid Algorithm:
• Compute ‘simple’ LK optical flow at hightest level
• At level i
• Take flow ui-1, vi-1 from level i-1
• Bilinear interpolate it to create ui
*, vi
* matrices of twice resolution for level I
• Multiply ui
*, vi
* by 2
• Warp level I Gaussian version of I2 according to predicted flow to create I2’
• Apply LK between I2’ and Gaussian version of I1 to get ui’(x, y), vi’(x, y)
• Add corrections ui’ vi’ i.e. ui = ui
* + vi
*
THANK YOU !
37

Weitere ähnliche Inhalte

Ähnlich wie Nhóm 13-OpticalFlow.pptx

Visual object tracking based on local steering kernals
Visual object tracking based on local steering kernalsVisual object tracking based on local steering kernals
Visual object tracking based on local steering kernalsSayahnarahul
 
Lec13 stereo converted
Lec13 stereo convertedLec13 stereo converted
Lec13 stereo convertedBaliThorat1
 
GPU-based Accelerated Spectral Caustic Rendering of Homogeneous Caustic Objects
GPU-based Accelerated Spectral Caustic Rendering of Homogeneous Caustic ObjectsGPU-based Accelerated Spectral Caustic Rendering of Homogeneous Caustic Objects
GPU-based Accelerated Spectral Caustic Rendering of Homogeneous Caustic ObjectsBudianto Tandianus
 
Computer Vision - Image Formation.pdf
Computer Vision - Image Formation.pdfComputer Vision - Image Formation.pdf
Computer Vision - Image Formation.pdfAmmarahMajeed
 
120_SEM_Special_Topics.ppt
120_SEM_Special_Topics.ppt120_SEM_Special_Topics.ppt
120_SEM_Special_Topics.pptzaki194502
 
Introduction to Binocular Stereo in Computer Vision
Introduction to Binocular Stereo in Computer VisionIntroduction to Binocular Stereo in Computer Vision
Introduction to Binocular Stereo in Computer Visionothersk46
 
Object detection at night
Object detection at nightObject detection at night
Object detection at nightSanjay Crúzé
 
Spectral clustering Tutorial
Spectral clustering TutorialSpectral clustering Tutorial
Spectral clustering TutorialZitao Liu
 
Quantum Computing Quantum Internet 2020_unit 1 By: Prof. Lili Saghafi
Quantum Computing Quantum Internet 2020_unit 1 By: Prof. Lili SaghafiQuantum Computing Quantum Internet 2020_unit 1 By: Prof. Lili Saghafi
Quantum Computing Quantum Internet 2020_unit 1 By: Prof. Lili SaghafiProfessor Lili Saghafi
 
Detection&Tracking - Thermal imaging object detection and tracking
Detection&Tracking - Thermal imaging object detection and trackingDetection&Tracking - Thermal imaging object detection and tracking
Detection&Tracking - Thermal imaging object detection and trackinghadarpinhas1
 
SPACE DEBRIS PALAT PARLAM no authors
SPACE DEBRIS PALAT PARLAM no authorsSPACE DEBRIS PALAT PARLAM no authors
SPACE DEBRIS PALAT PARLAM no authorsMARIUS EUGEN OPRAN
 
Solar System Processing with LSST: A Status Update
Solar System Processing with LSST: A Status UpdateSolar System Processing with LSST: A Status Update
Solar System Processing with LSST: A Status UpdateMario Juric
 
Temporal Superpixels Based on Proximity-Weighted Patch Matching
Temporal Superpixels Based on Proximity-Weighted Patch MatchingTemporal Superpixels Based on Proximity-Weighted Patch Matching
Temporal Superpixels Based on Proximity-Weighted Patch MatchingNAVER Engineering
 
Intro to Neural Networks
Intro to Neural NetworksIntro to Neural Networks
Intro to Neural NetworksDean Wyatte
 
Paris Master Class 2011 - 07 Dynamic Global Illumination
Paris Master Class 2011 - 07 Dynamic Global IlluminationParis Master Class 2011 - 07 Dynamic Global Illumination
Paris Master Class 2011 - 07 Dynamic Global IlluminationWolfgang Engel
 
Oxford 05-oct-2012
Oxford 05-oct-2012Oxford 05-oct-2012
Oxford 05-oct-2012Ted Dunning
 
PR-132: SSD: Single Shot MultiBox Detector
PR-132: SSD: Single Shot MultiBox DetectorPR-132: SSD: Single Shot MultiBox Detector
PR-132: SSD: Single Shot MultiBox DetectorJinwon Lee
 

Ähnlich wie Nhóm 13-OpticalFlow.pptx (20)

Visual object tracking based on local steering kernals
Visual object tracking based on local steering kernalsVisual object tracking based on local steering kernals
Visual object tracking based on local steering kernals
 
Lec13 stereo converted
Lec13 stereo convertedLec13 stereo converted
Lec13 stereo converted
 
GPU-based Accelerated Spectral Caustic Rendering of Homogeneous Caustic Objects
GPU-based Accelerated Spectral Caustic Rendering of Homogeneous Caustic ObjectsGPU-based Accelerated Spectral Caustic Rendering of Homogeneous Caustic Objects
GPU-based Accelerated Spectral Caustic Rendering of Homogeneous Caustic Objects
 
Computer Vision - Image Formation.pdf
Computer Vision - Image Formation.pdfComputer Vision - Image Formation.pdf
Computer Vision - Image Formation.pdf
 
120_SEM_Special_Topics.ppt
120_SEM_Special_Topics.ppt120_SEM_Special_Topics.ppt
120_SEM_Special_Topics.ppt
 
Introduction to Binocular Stereo in Computer Vision
Introduction to Binocular Stereo in Computer VisionIntroduction to Binocular Stereo in Computer Vision
Introduction to Binocular Stereo in Computer Vision
 
Object detection at night
Object detection at nightObject detection at night
Object detection at night
 
november6.ppt
november6.pptnovember6.ppt
november6.ppt
 
Spectral clustering Tutorial
Spectral clustering TutorialSpectral clustering Tutorial
Spectral clustering Tutorial
 
Quantum Computing Quantum Internet 2020_unit 1 By: Prof. Lili Saghafi
Quantum Computing Quantum Internet 2020_unit 1 By: Prof. Lili SaghafiQuantum Computing Quantum Internet 2020_unit 1 By: Prof. Lili Saghafi
Quantum Computing Quantum Internet 2020_unit 1 By: Prof. Lili Saghafi
 
Detection&Tracking - Thermal imaging object detection and tracking
Detection&Tracking - Thermal imaging object detection and trackingDetection&Tracking - Thermal imaging object detection and tracking
Detection&Tracking - Thermal imaging object detection and tracking
 
SPACE DEBRIS PALAT PARLAM no authors
SPACE DEBRIS PALAT PARLAM no authorsSPACE DEBRIS PALAT PARLAM no authors
SPACE DEBRIS PALAT PARLAM no authors
 
Prelim
PrelimPrelim
Prelim
 
Solar System Processing with LSST: A Status Update
Solar System Processing with LSST: A Status UpdateSolar System Processing with LSST: A Status Update
Solar System Processing with LSST: A Status Update
 
Temporal Superpixels Based on Proximity-Weighted Patch Matching
Temporal Superpixels Based on Proximity-Weighted Patch MatchingTemporal Superpixels Based on Proximity-Weighted Patch Matching
Temporal Superpixels Based on Proximity-Weighted Patch Matching
 
Time lens
Time lensTime lens
Time lens
 
Intro to Neural Networks
Intro to Neural NetworksIntro to Neural Networks
Intro to Neural Networks
 
Paris Master Class 2011 - 07 Dynamic Global Illumination
Paris Master Class 2011 - 07 Dynamic Global IlluminationParis Master Class 2011 - 07 Dynamic Global Illumination
Paris Master Class 2011 - 07 Dynamic Global Illumination
 
Oxford 05-oct-2012
Oxford 05-oct-2012Oxford 05-oct-2012
Oxford 05-oct-2012
 
PR-132: SSD: Single Shot MultiBox Detector
PR-132: SSD: Single Shot MultiBox DetectorPR-132: SSD: Single Shot MultiBox Detector
PR-132: SSD: Single Shot MultiBox Detector
 

Kürzlich hochgeladen

Science 7 Quarter 4 Module 2: Natural Resources.pptx
Science 7 Quarter 4 Module 2: Natural Resources.pptxScience 7 Quarter 4 Module 2: Natural Resources.pptx
Science 7 Quarter 4 Module 2: Natural Resources.pptxMaryGraceBautista27
 
4.18.24 Movement Legacies, Reflection, and Review.pptx
4.18.24 Movement Legacies, Reflection, and Review.pptx4.18.24 Movement Legacies, Reflection, and Review.pptx
4.18.24 Movement Legacies, Reflection, and Review.pptxmary850239
 
HỌC TỐT TIẾNG ANH 11 THEO CHƯƠNG TRÌNH GLOBAL SUCCESS ĐÁP ÁN CHI TIẾT - CẢ NĂ...
HỌC TỐT TIẾNG ANH 11 THEO CHƯƠNG TRÌNH GLOBAL SUCCESS ĐÁP ÁN CHI TIẾT - CẢ NĂ...HỌC TỐT TIẾNG ANH 11 THEO CHƯƠNG TRÌNH GLOBAL SUCCESS ĐÁP ÁN CHI TIẾT - CẢ NĂ...
HỌC TỐT TIẾNG ANH 11 THEO CHƯƠNG TRÌNH GLOBAL SUCCESS ĐÁP ÁN CHI TIẾT - CẢ NĂ...Nguyen Thanh Tu Collection
 
Difference Between Search & Browse Methods in Odoo 17
Difference Between Search & Browse Methods in Odoo 17Difference Between Search & Browse Methods in Odoo 17
Difference Between Search & Browse Methods in Odoo 17Celine George
 
ANG SEKTOR NG agrikultura.pptx QUARTER 4
ANG SEKTOR NG agrikultura.pptx QUARTER 4ANG SEKTOR NG agrikultura.pptx QUARTER 4
ANG SEKTOR NG agrikultura.pptx QUARTER 4MiaBumagat1
 
Concurrency Control in Database Management system
Concurrency Control in Database Management systemConcurrency Control in Database Management system
Concurrency Control in Database Management systemChristalin Nelson
 
AUDIENCE THEORY -CULTIVATION THEORY - GERBNER.pptx
AUDIENCE THEORY -CULTIVATION THEORY -  GERBNER.pptxAUDIENCE THEORY -CULTIVATION THEORY -  GERBNER.pptx
AUDIENCE THEORY -CULTIVATION THEORY - GERBNER.pptxiammrhaywood
 
Visit to a blind student's school🧑‍🦯🧑‍🦯(community medicine)
Visit to a blind student's school🧑‍🦯🧑‍🦯(community medicine)Visit to a blind student's school🧑‍🦯🧑‍🦯(community medicine)
Visit to a blind student's school🧑‍🦯🧑‍🦯(community medicine)lakshayb543
 
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️9953056974 Low Rate Call Girls In Saket, Delhi NCR
 
Field Attribute Index Feature in Odoo 17
Field Attribute Index Feature in Odoo 17Field Attribute Index Feature in Odoo 17
Field Attribute Index Feature in Odoo 17Celine George
 
Transaction Management in Database Management System
Transaction Management in Database Management SystemTransaction Management in Database Management System
Transaction Management in Database Management SystemChristalin Nelson
 
Procuring digital preservation CAN be quick and painless with our new dynamic...
Procuring digital preservation CAN be quick and painless with our new dynamic...Procuring digital preservation CAN be quick and painless with our new dynamic...
Procuring digital preservation CAN be quick and painless with our new dynamic...Jisc
 
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTS
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTSGRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTS
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTSJoshuaGantuangco2
 
ENGLISH6-Q4-W3.pptxqurter our high choom
ENGLISH6-Q4-W3.pptxqurter our high choomENGLISH6-Q4-W3.pptxqurter our high choom
ENGLISH6-Q4-W3.pptxqurter our high choomnelietumpap1
 
Like-prefer-love -hate+verb+ing & silent letters & citizenship text.pdf
Like-prefer-love -hate+verb+ing & silent letters & citizenship text.pdfLike-prefer-love -hate+verb+ing & silent letters & citizenship text.pdf
Like-prefer-love -hate+verb+ing & silent letters & citizenship text.pdfMr Bounab Samir
 
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...JhezDiaz1
 

Kürzlich hochgeladen (20)

Science 7 Quarter 4 Module 2: Natural Resources.pptx
Science 7 Quarter 4 Module 2: Natural Resources.pptxScience 7 Quarter 4 Module 2: Natural Resources.pptx
Science 7 Quarter 4 Module 2: Natural Resources.pptx
 
4.18.24 Movement Legacies, Reflection, and Review.pptx
4.18.24 Movement Legacies, Reflection, and Review.pptx4.18.24 Movement Legacies, Reflection, and Review.pptx
4.18.24 Movement Legacies, Reflection, and Review.pptx
 
HỌC TỐT TIẾNG ANH 11 THEO CHƯƠNG TRÌNH GLOBAL SUCCESS ĐÁP ÁN CHI TIẾT - CẢ NĂ...
HỌC TỐT TIẾNG ANH 11 THEO CHƯƠNG TRÌNH GLOBAL SUCCESS ĐÁP ÁN CHI TIẾT - CẢ NĂ...HỌC TỐT TIẾNG ANH 11 THEO CHƯƠNG TRÌNH GLOBAL SUCCESS ĐÁP ÁN CHI TIẾT - CẢ NĂ...
HỌC TỐT TIẾNG ANH 11 THEO CHƯƠNG TRÌNH GLOBAL SUCCESS ĐÁP ÁN CHI TIẾT - CẢ NĂ...
 
Difference Between Search & Browse Methods in Odoo 17
Difference Between Search & Browse Methods in Odoo 17Difference Between Search & Browse Methods in Odoo 17
Difference Between Search & Browse Methods in Odoo 17
 
YOUVE_GOT_EMAIL_PRELIMS_EL_DORADO_2024.pptx
YOUVE_GOT_EMAIL_PRELIMS_EL_DORADO_2024.pptxYOUVE_GOT_EMAIL_PRELIMS_EL_DORADO_2024.pptx
YOUVE_GOT_EMAIL_PRELIMS_EL_DORADO_2024.pptx
 
ANG SEKTOR NG agrikultura.pptx QUARTER 4
ANG SEKTOR NG agrikultura.pptx QUARTER 4ANG SEKTOR NG agrikultura.pptx QUARTER 4
ANG SEKTOR NG agrikultura.pptx QUARTER 4
 
Concurrency Control in Database Management system
Concurrency Control in Database Management systemConcurrency Control in Database Management system
Concurrency Control in Database Management system
 
AUDIENCE THEORY -CULTIVATION THEORY - GERBNER.pptx
AUDIENCE THEORY -CULTIVATION THEORY -  GERBNER.pptxAUDIENCE THEORY -CULTIVATION THEORY -  GERBNER.pptx
AUDIENCE THEORY -CULTIVATION THEORY - GERBNER.pptx
 
Visit to a blind student's school🧑‍🦯🧑‍🦯(community medicine)
Visit to a blind student's school🧑‍🦯🧑‍🦯(community medicine)Visit to a blind student's school🧑‍🦯🧑‍🦯(community medicine)
Visit to a blind student's school🧑‍🦯🧑‍🦯(community medicine)
 
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️
 
LEFT_ON_C'N_ PRELIMS_EL_DORADO_2024.pptx
LEFT_ON_C'N_ PRELIMS_EL_DORADO_2024.pptxLEFT_ON_C'N_ PRELIMS_EL_DORADO_2024.pptx
LEFT_ON_C'N_ PRELIMS_EL_DORADO_2024.pptx
 
Field Attribute Index Feature in Odoo 17
Field Attribute Index Feature in Odoo 17Field Attribute Index Feature in Odoo 17
Field Attribute Index Feature in Odoo 17
 
Transaction Management in Database Management System
Transaction Management in Database Management SystemTransaction Management in Database Management System
Transaction Management in Database Management System
 
Raw materials used in Herbal Cosmetics.pptx
Raw materials used in Herbal Cosmetics.pptxRaw materials used in Herbal Cosmetics.pptx
Raw materials used in Herbal Cosmetics.pptx
 
Procuring digital preservation CAN be quick and painless with our new dynamic...
Procuring digital preservation CAN be quick and painless with our new dynamic...Procuring digital preservation CAN be quick and painless with our new dynamic...
Procuring digital preservation CAN be quick and painless with our new dynamic...
 
FINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptx
FINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptxFINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptx
FINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptx
 
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTS
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTSGRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTS
GRADE 4 - SUMMATIVE TEST QUARTER 4 ALL SUBJECTS
 
ENGLISH6-Q4-W3.pptxqurter our high choom
ENGLISH6-Q4-W3.pptxqurter our high choomENGLISH6-Q4-W3.pptxqurter our high choom
ENGLISH6-Q4-W3.pptxqurter our high choom
 
Like-prefer-love -hate+verb+ing & silent letters & citizenship text.pdf
Like-prefer-love -hate+verb+ing & silent letters & citizenship text.pdfLike-prefer-love -hate+verb+ing & silent letters & citizenship text.pdf
Like-prefer-love -hate+verb+ing & silent letters & citizenship text.pdf
 
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...
 

Nhóm 13-OpticalFlow.pptx

  • 1. OPTICAL FLOW Team 13 Phạm Quang Anh Nguyễn Đức Quang Nguyễn Quang Anh Đỗ Văn Anh 1
  • 2. Table of contents 1. Problem Definition 2. Types of Optical Flow 3. Sparse Optical Flow – Lucas-Kanade Algorithm 4. Dense Optical Flow – Horn-Schunck Algorithm 5. Conclusion 2
  • 4. 1. Problem Definition • Optical Flow: Optical flow is the motion of objects between consecutive frames of sequence, caused by the relative movement between the object and camera. • Problem: Given two consecutive image frames, estimate the motion of each pixel. 4
  • 7. 2. Types of Optical Flow • Sparse Optical Flow: Sparse optical flow gives the flow vectors of some "interesting features" (say few pixels depicting the edges or corners of an object) within the frame. 7
  • 8. 2. Types of Optical Flow • Dense Optical Flow: Dense optical flow attempts to compute the optical flow vector for every pixel of each frame. • While such computation may be slower, it gives a more accurate result and a denser result suitable for applications such as learning structure from motion and video segmentation. 8
  • 10. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Lucas and Kanade proposed an effective technique to estimate the motion of interesting features by comparing two consecutive frames. • The Lucas-Kanade method works under the brightness constancy assumption and small motion assumption. 10
  • 11. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Shi-Tomashi Corner Detection: For the implementation of sparse optical flow, we only track the motion of a feature set of pixels. Features in images are points of interest which present rich image content information. 11
  • 12. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Shi-Tomashi Corner Detection: 12
  • 13. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Shi-Tomashi Corner Detection: 13
  • 14. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Shi-Tomashi Corner Detection: 14
  • 15. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Brightness constancy assumption: Brightness of the point will remain the same. I(x(t), y(t), t) = C - constant 15
  • 16. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Small motion assumption: 16
  • 17. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Small motion assumption: 17
  • 18. 3. Sparse Optical Flow – Lucas-Kanade Algorithm 18
  • 19. 3. Sparse Optical Flow – Lucas-Kanade Algorithm 19
  • 20. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Aperture problem: 20
  • 21. 3. Sparse Optical Flow – Lucas-Kanade Algorithm 21
  • 22. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Smoothness constancy assumption: A frame portrays a “natural” scene with textured objects exhibiting shades of gray that change smoothly. 22
  • 23. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Smoothness constancy assumption: 23
  • 24. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Smoothness constancy assumption: 24
  • 25. 3. Sparse Optical Flow – Lucas-Kanade Algorithm • Smoothness constancy assumption: 25
  • 27. 4. Dense Optical Flow – Horn-Schunck Algorithm • The Horn–Schunck method of estimating optical flow is a global method which introduces a global constraint of smoothness to solve the aperture problem. 27
  • 28. 4. Dense Optical Flow – Horn-Schunck Algorithm 28
  • 29. 4. Dense Optical Flow – Horn-Schunck Algorithm 29
  • 30. 4. Dense Optical Flow – Horn-Schunck Algorithm 30
  • 31. 4. Dense Optical Flow – Horn-Schunck Algorithm 31
  • 33. 5. Lucas Kanade with Pyramid 33 In regular optical flow method, we assume the following: • Brightness constancy • Small motion • Spatial conherence If the object were to move a larger distance →The traditional optical flow method would work bad
  • 34. 5. Lucas Kanade with Pyramid 34 • Pyramid is built by using multiple copies of the same image. • Each level in the pyramid is 1/4th of the size of the previous level • The lowest level is of the highest resolution • The highest level is of the lowest resolution
  • 35. 5. Lucas Kanade with Pyramid 35 • Pyramid is built by using multiple copies of the same image. • Each level in the pyramid is 1/4th of the size of the previous level • The lowest level is of the highest resolution • The highest level is of the lowest resolution • To Downsample: Using Gausian pyramid • To Upsamgple: Using Laplacian pyramid
  • 36. 5. Lucas Kanade with Pyramid 36 Lucas-Kanade with Pyramid Algorithm: • Compute ‘simple’ LK optical flow at hightest level • At level i • Take flow ui-1, vi-1 from level i-1 • Bilinear interpolate it to create ui *, vi * matrices of twice resolution for level I • Multiply ui *, vi * by 2 • Warp level I Gaussian version of I2 according to predicted flow to create I2’ • Apply LK between I2’ and Gaussian version of I1 to get ui’(x, y), vi’(x, y) • Add corrections ui’ vi’ i.e. ui = ui * + vi *