SlideShare une entreprise Scribd logo
1  sur  14
Normalized averaging
      using adaptive applicability functions

                            with applications in image reconstruction
                            from sparsely and randomly sampled data




      Presented at SCIA 2003

      Tuan Q. Pham and Lucas J. van Vliet

      July 17, 2009

                                                                        1




Pattern Recognition Group
Overview


• Normalized averaging
• Local structure adaptive filtering
• Experimental results
• Comparison with diffusion-based image inpainting
• Directions for further research




July 17, 2009                                        2
Normalized averaging (Knutsson’93)
•Weighted average filtering:                                                 r = s*a

•Normalized averaging = weighted average + signal/certainty principle:
        •each signal s is associated with a certainty c
        •s & c have to be processed separately                               ( s . c) ∗ a
                                                                          r=
                                                                                c∗a
                where s :signal, c :certainty, a :filter, r :result, * :convolution




                input with 10%          Gaussian smoothing          NA with Gaussian
                 original pixels              (σ = 1)              applicability (σ = 1)
July 17, 2009                                                                               3
Normalized averaging: An example
Reconstruction from 10% random pixels




Nearest neighbor interpolation   NA with adaptive applicability
July 17, 2009                                                     4
Image reconstruction using
Adaptive Normalized Averaging


   Input Image   σ=1     Normalized    Output Image
   (sparsely &           Averaging     (with local
   randomly                            structure
   sampled)            Adaptive        extended into
                       applicability   missing regions)

                          Structure
                          Analysis


July 17, 2009                                             5
Local structure adaptive filtering
•Local structure from the structure tensor                                          r
                                                                                    v
                                                                            r
                      r r T                                                 u ϕ
                                  rr T  rr T
                T = ∇I ∇I = λu uu + λv vv
                                      r
                •orientation φ = arg(u)
                                                                       y = 1 κ x2
                •anisotropy A = (λu - λv)/(λu + λv)                        2
                                        r
                •curvature    κ = ∂φ /∂ v
                •scale    rdensity = sample density

•Scale-adaptive curvature-bent anisotropic Gaussian
kernel with scales in 2 orthogonal directions:
     σ u = C (1 − A)α rdensity σ v = C (1 + A)α rdensity
where           C ~ SNR       α   ~ degree of structure enhancement   kernel aligns with
                                                                       local structure
July 17, 2009                                                                           6
Sample Density Transform
     •Definition: Smallest radius of a pillbox, centered at each pixel, that
     encompasses total certainty of at least 1

     •Role: Automatic scale selection of the applicability in the NA equation
     to avoid unnecessary smoothing




Lena with missing hole   Density transform   NA with Gaussian(σ=1)   Adap. Norm. Avg.

     July 17, 2009                                                              7
4x4 super-resolution from 4 noisy frames
•     4 input LowRes captured with fill-factor = 25%, intensity noise
      (σ=10), registration noise (σ=0.2 LR pitch)




1 of 4 input 64x64 LR         SR using triangulation   SR using adaptive NA
• 16 times upsampling from only 4 frames. How is it possible: along
  linear structures, only 4 samples are enough for 4x super-resolution
 July 17, 2009                                                          8
Orientation            Anisotropy                     Curvature




Sample density   Scale along linear structures Scale in perpendicular direction
July 17, 2009                                                         9
Comparison with image inpainting

• Image inpainting (Sapiro) = diffusion with level line evolution
        • also extending orientation into the missing regions
        • slow due to iterative nature
        • poor result for large holes




     input                inpainting          inpainting +      Adapt. Norm. Avg.
                       110 iters (6 min)    texture synthesis     0 iters (6 sec)
July 17, 2009                                                                10
Directions for Further Research

• Applications
        • Image filtering (noise/watermark removal, edge enhancement...)
        • Image interpolation from sparsely and randomly sampled data
          (image inpainting, image fusion, super-resolution...)


• Further improvements
        • Scale-space local structure analysis.
        • Detect multiple orientations using orientation space.
        • Robust neighborhood operator than the weighted mean.




July 17, 2009                                                      11
Image inpainting of thin scribbles




       input




  inpainting     Adaptive Normalized Averaging (10 sec)
July 17, 2009                                         12
Simultaneous geometry/texture inpainting




                                        texture
  input




geometry                               Adaptive NA
                                         (1 min)

    July 17, 2009                            13
Inpainting of ambiguous discontinuity




      original   input   inpainting   Adaptive NA (1 sec)




July 17, 2009                                        14

Contenu connexe

Tendances

study Image and video abstraction by multi scale anisotropic kuwahara
study  Image and video abstraction by multi scale anisotropic kuwaharastudy  Image and video abstraction by multi scale anisotropic kuwahara
study Image and video abstraction by multi scale anisotropic kuwahara
Chiamin Hsu
 
Learning Moving Cast Shadows for Foreground Detection (VS 2008)
Learning Moving Cast Shadows for Foreground Detection (VS 2008)Learning Moving Cast Shadows for Foreground Detection (VS 2008)
Learning Moving Cast Shadows for Foreground Detection (VS 2008)
Jia-Bin Huang
 
Pisek_20110726_IGARSS2011_v2.ppt
Pisek_20110726_IGARSS2011_v2.pptPisek_20110726_IGARSS2011_v2.ppt
Pisek_20110726_IGARSS2011_v2.ppt
grssieee
 

Tendances (6)

Robust Shape and Topology Optimization - Northwestern
Robust Shape and Topology Optimization - Northwestern Robust Shape and Topology Optimization - Northwestern
Robust Shape and Topology Optimization - Northwestern
 
study Image and video abstraction by multi scale anisotropic kuwahara
study  Image and video abstraction by multi scale anisotropic kuwaharastudy  Image and video abstraction by multi scale anisotropic kuwahara
study Image and video abstraction by multi scale anisotropic kuwahara
 
Learning Moving Cast Shadows for Foreground Detection (VS 2008)
Learning Moving Cast Shadows for Foreground Detection (VS 2008)Learning Moving Cast Shadows for Foreground Detection (VS 2008)
Learning Moving Cast Shadows for Foreground Detection (VS 2008)
 
Pisek_20110726_IGARSS2011_v2.ppt
Pisek_20110726_IGARSS2011_v2.pptPisek_20110726_IGARSS2011_v2.ppt
Pisek_20110726_IGARSS2011_v2.ppt
 
Topological Inference via Meshing
Topological Inference via MeshingTopological Inference via Meshing
Topological Inference via Meshing
 
Ostst 2014 - scagliola
Ostst 2014 - scagliolaOstst 2014 - scagliola
Ostst 2014 - scagliola
 

En vedette

Assignment2 analog to digital conversion soumit_mukherjee
Assignment2 analog to digital conversion soumit_mukherjeeAssignment2 analog to digital conversion soumit_mukherjee
Assignment2 analog to digital conversion soumit_mukherjee
Soumit Mukherjee
 
Kalman filter - Applications in Image processing
Kalman filter - Applications in Image processingKalman filter - Applications in Image processing
Kalman filter - Applications in Image processing
Ravi Teja
 
The Low Vision Examination
The Low Vision ExaminationThe Low Vision Examination
The Low Vision Examination
Hossein Mirzaie
 
Sensors On 3d Digitization
Sensors On 3d DigitizationSensors On 3d Digitization
Sensors On 3d Digitization
Rajan Kumar
 
Digital image processing img smoothning
Digital image processing img smoothningDigital image processing img smoothning
Digital image processing img smoothning
Vinay Gupta
 

En vedette (15)

Super resolution microscopy & astronomy
Super resolution microscopy & astronomySuper resolution microscopy & astronomy
Super resolution microscopy & astronomy
 
Icdecs 2011
Icdecs 2011Icdecs 2011
Icdecs 2011
 
Assignment2 analog to digital conversion soumit_mukherjee
Assignment2 analog to digital conversion soumit_mukherjeeAssignment2 analog to digital conversion soumit_mukherjee
Assignment2 analog to digital conversion soumit_mukherjee
 
Kalman filter - Applications in Image processing
Kalman filter - Applications in Image processingKalman filter - Applications in Image processing
Kalman filter - Applications in Image processing
 
Retina preliminary
Retina  preliminaryRetina  preliminary
Retina preliminary
 
Single Image Super-Resolution from Transformed Self-Exemplars (CVPR 2015)
Single Image Super-Resolution from Transformed Self-Exemplars (CVPR 2015)Single Image Super-Resolution from Transformed Self-Exemplars (CVPR 2015)
Single Image Super-Resolution from Transformed Self-Exemplars (CVPR 2015)
 
Data compression
Data compression Data compression
Data compression
 
Image processing fundamentals
Image processing fundamentalsImage processing fundamentals
Image processing fundamentals
 
The Low Vision Examination
The Low Vision ExaminationThe Low Vision Examination
The Low Vision Examination
 
Matlab Image Restoration Techniques
Matlab Image Restoration TechniquesMatlab Image Restoration Techniques
Matlab Image Restoration Techniques
 
Sensors On 3d Digitization
Sensors On 3d DigitizationSensors On 3d Digitization
Sensors On 3d Digitization
 
Image processing
Image processingImage processing
Image processing
 
Bio-metrics Technology
Bio-metrics TechnologyBio-metrics Technology
Bio-metrics Technology
 
Data Compression (Lossy and Lossless)
Data Compression (Lossy and Lossless)Data Compression (Lossy and Lossless)
Data Compression (Lossy and Lossless)
 
Digital image processing img smoothning
Digital image processing img smoothningDigital image processing img smoothning
Digital image processing img smoothning
 

Similaire à Normalized averaging using adaptive applicability functions with applications in image reconstruction from sparsely and randomly sampled data

Realism of image composits
Realism of image compositsRealism of image composits
Realism of image composits
Aliya Ibragimova
 
ADAPTIVE LOCAL KRIGING (ALK) TO RETRIEVE THE SLANT RANGE SURFACE MOTION MAPS ...
ADAPTIVE LOCAL KRIGING (ALK) TO RETRIEVE THE SLANT RANGE SURFACE MOTION MAPS ...ADAPTIVE LOCAL KRIGING (ALK) TO RETRIEVE THE SLANT RANGE SURFACE MOTION MAPS ...
ADAPTIVE LOCAL KRIGING (ALK) TO RETRIEVE THE SLANT RANGE SURFACE MOTION MAPS ...
grssieee
 
Lecture04
Lecture04Lecture04
Lecture04
zukun
 
Image Smoothing for Structure Extraction
Image Smoothing for Structure ExtractionImage Smoothing for Structure Extraction
Image Smoothing for Structure Extraction
Jia-Bin Huang
 
Presentation at SMI 2023
Presentation at SMI 2023Presentation at SMI 2023
Presentation at SMI 2023
Joaquim Jorge
 
Fingerprint _prem
Fingerprint _premFingerprint _prem
Fingerprint _prem
lgbl40
 
Andrew Zisserman Talk - Part 1a
Andrew Zisserman Talk - Part 1aAndrew Zisserman Talk - Part 1a
Andrew Zisserman Talk - Part 1a
Anton Konushin
 
Accelarating Optical Quadrature Microscopy Using GPUs
Accelarating Optical Quadrature Microscopy Using GPUsAccelarating Optical Quadrature Microscopy Using GPUs
Accelarating Optical Quadrature Microscopy Using GPUs
Perhaad Mistry
 
Xu_IGARSS11_snow_f.pdf
Xu_IGARSS11_snow_f.pdfXu_IGARSS11_snow_f.pdf
Xu_IGARSS11_snow_f.pdf
grssieee
 
Biao Hou--SAR IMAGE DESPECKLING BASED ON IMPROVED DIRECTIONLET DOMAIN GAUSSIA...
Biao Hou--SAR IMAGE DESPECKLING BASED ON IMPROVED DIRECTIONLET DOMAIN GAUSSIA...Biao Hou--SAR IMAGE DESPECKLING BASED ON IMPROVED DIRECTIONLET DOMAIN GAUSSIA...
Biao Hou--SAR IMAGE DESPECKLING BASED ON IMPROVED DIRECTIONLET DOMAIN GAUSSIA...
grssieee
 

Similaire à Normalized averaging using adaptive applicability functions with applications in image reconstruction from sparsely and randomly sampled data (20)

Realism of image composits
Realism of image compositsRealism of image composits
Realism of image composits
 
Surveys
SurveysSurveys
Surveys
 
ADAPTIVE LOCAL KRIGING (ALK) TO RETRIEVE THE SLANT RANGE SURFACE MOTION MAPS ...
ADAPTIVE LOCAL KRIGING (ALK) TO RETRIEVE THE SLANT RANGE SURFACE MOTION MAPS ...ADAPTIVE LOCAL KRIGING (ALK) TO RETRIEVE THE SLANT RANGE SURFACE MOTION MAPS ...
ADAPTIVE LOCAL KRIGING (ALK) TO RETRIEVE THE SLANT RANGE SURFACE MOTION MAPS ...
 
Lecture04
Lecture04Lecture04
Lecture04
 
Image Smoothing for Structure Extraction
Image Smoothing for Structure ExtractionImage Smoothing for Structure Extraction
Image Smoothing for Structure Extraction
 
Talk
TalkTalk
Talk
 
Presentation at SMI 2023
Presentation at SMI 2023Presentation at SMI 2023
Presentation at SMI 2023
 
Sparse and Redundant Representations: Theory and Applications
Sparse and Redundant Representations: Theory and ApplicationsSparse and Redundant Representations: Theory and Applications
Sparse and Redundant Representations: Theory and Applications
 
Normalized averaging using adaptive applicability functions with applications...
Normalized averaging using adaptive applicability functions with applications...Normalized averaging using adaptive applicability functions with applications...
Normalized averaging using adaptive applicability functions with applications...
 
Fingerprint _prem
Fingerprint _premFingerprint _prem
Fingerprint _prem
 
Andrew Zisserman Talk - Part 1a
Andrew Zisserman Talk - Part 1aAndrew Zisserman Talk - Part 1a
Andrew Zisserman Talk - Part 1a
 
Accelarating Optical Quadrature Microscopy Using GPUs
Accelarating Optical Quadrature Microscopy Using GPUsAccelarating Optical Quadrature Microscopy Using GPUs
Accelarating Optical Quadrature Microscopy Using GPUs
 
UCB 2012-02-28
UCB 2012-02-28UCB 2012-02-28
UCB 2012-02-28
 
Xu_IGARSS11_snow_f.pdf
Xu_IGARSS11_snow_f.pdfXu_IGARSS11_snow_f.pdf
Xu_IGARSS11_snow_f.pdf
 
Chapter10 image segmentation
Chapter10 image segmentationChapter10 image segmentation
Chapter10 image segmentation
 
190 195
190 195190 195
190 195
 
Scale Invariant feature transform
Scale Invariant feature transformScale Invariant feature transform
Scale Invariant feature transform
 
Biao Hou--SAR IMAGE DESPECKLING BASED ON IMPROVED DIRECTIONLET DOMAIN GAUSSIA...
Biao Hou--SAR IMAGE DESPECKLING BASED ON IMPROVED DIRECTIONLET DOMAIN GAUSSIA...Biao Hou--SAR IMAGE DESPECKLING BASED ON IMPROVED DIRECTIONLET DOMAIN GAUSSIA...
Biao Hou--SAR IMAGE DESPECKLING BASED ON IMPROVED DIRECTIONLET DOMAIN GAUSSIA...
 
Elementary Landscape Decomposition of the Test Suite Minimization Problem
Elementary Landscape Decomposition of the Test Suite Minimization ProblemElementary Landscape Decomposition of the Test Suite Minimization Problem
Elementary Landscape Decomposition of the Test Suite Minimization Problem
 
Edges and lines
Edges and linesEdges and lines
Edges and lines
 

Plus de Tuan Q. Pham

Asymmetric recursive Gaussian filtering for space-variant artificial bokeh
 Asymmetric recursive Gaussian filtering for space-variant artificial bokeh Asymmetric recursive Gaussian filtering for space-variant artificial bokeh
Asymmetric recursive Gaussian filtering for space-variant artificial bokeh
Tuan Q. Pham
 

Plus de Tuan Q. Pham (15)

Oral presentation on Asymmetric recursive Gaussian filtering for space-varia...
Oral presentation on  Asymmetric recursive Gaussian filtering for space-varia...Oral presentation on  Asymmetric recursive Gaussian filtering for space-varia...
Oral presentation on Asymmetric recursive Gaussian filtering for space-varia...
 
Asymmetric recursive Gaussian filtering for space-variant artificial bokeh
 Asymmetric recursive Gaussian filtering for space-variant artificial bokeh Asymmetric recursive Gaussian filtering for space-variant artificial bokeh
Asymmetric recursive Gaussian filtering for space-variant artificial bokeh
 
Parallel implementation of geodesic distance transform with application in su...
Parallel implementation of geodesic distance transform with application in su...Parallel implementation of geodesic distance transform with application in su...
Parallel implementation of geodesic distance transform with application in su...
 
Parallel implementation of geodesic distance transform with application in su...
Parallel implementation of geodesic distance transform with application in su...Parallel implementation of geodesic distance transform with application in su...
Parallel implementation of geodesic distance transform with application in su...
 
Multi-hypothesis projection-based shift estimation for sweeping panorama reco...
Multi-hypothesis projection-based shift estimation for sweeping panorama reco...Multi-hypothesis projection-based shift estimation for sweeping panorama reco...
Multi-hypothesis projection-based shift estimation for sweeping panorama reco...
 
Multi-hypothesis projection-based shift estimation for sweeping panorama reco...
Multi-hypothesis projection-based shift estimation for sweeping panorama reco...Multi-hypothesis projection-based shift estimation for sweeping panorama reco...
Multi-hypothesis projection-based shift estimation for sweeping panorama reco...
 
Non-maximum suppression using fewer than two comparison per pixels
Non-maximum suppression using fewer than two comparison per pixelsNon-maximum suppression using fewer than two comparison per pixels
Non-maximum suppression using fewer than two comparison per pixels
 
Paper fingerprinting using alpha-masked image matching
Paper fingerprinting using alpha-masked image matchingPaper fingerprinting using alpha-masked image matching
Paper fingerprinting using alpha-masked image matching
 
Paper fingerprinting using alpha-masked image matching
Paper fingerprinting using alpha-masked image matchingPaper fingerprinting using alpha-masked image matching
Paper fingerprinting using alpha-masked image matching
 
Bidirectional bias correction for gradient-based shift estimation
Bidirectional bias correction for gradient-based shift estimationBidirectional bias correction for gradient-based shift estimation
Bidirectional bias correction for gradient-based shift estimation
 
Robust Super-Resolution by minimizing a Gaussian-weighted L2 error norm
Robust Super-Resolution by minimizing a Gaussian-weighted L2 error normRobust Super-Resolution by minimizing a Gaussian-weighted L2 error norm
Robust Super-Resolution by minimizing a Gaussian-weighted L2 error norm
 
Resolution enhancement of low-quality videos using a high-resolution frame
Resolution enhancement of low-quality videos using a high-resolution frameResolution enhancement of low-quality videos using a high-resolution frame
Resolution enhancement of low-quality videos using a high-resolution frame
 
Separable bilateral filtering for fast video preprocessing
Separable bilateral filtering for fast video preprocessingSeparable bilateral filtering for fast video preprocessing
Separable bilateral filtering for fast video preprocessing
 
Performance of Optimal Registration Estimator
Performance of Optimal Registration EstimatorPerformance of Optimal Registration Estimator
Performance of Optimal Registration Estimator
 
Influence of Signal-to-Noise Ratio and Point Spread Function on Limits of Sup...
Influence of Signal-to-Noise Ratio and Point Spread Function on Limits of Sup...Influence of Signal-to-Noise Ratio and Point Spread Function on Limits of Sup...
Influence of Signal-to-Noise Ratio and Point Spread Function on Limits of Sup...
 

Dernier

Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
Victor Rentea
 
Architecting Cloud Native Applications
Architecting Cloud Native ApplicationsArchitecting Cloud Native Applications
Architecting Cloud Native Applications
WSO2
 

Dernier (20)

DBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor PresentationDBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor Presentation
 
Artificial Intelligence Chap.5 : Uncertainty
Artificial Intelligence Chap.5 : UncertaintyArtificial Intelligence Chap.5 : Uncertainty
Artificial Intelligence Chap.5 : Uncertainty
 
CNIC Information System with Pakdata Cf In Pakistan
CNIC Information System with Pakdata Cf In PakistanCNIC Information System with Pakdata Cf In Pakistan
CNIC Information System with Pakdata Cf In Pakistan
 
Strategize a Smooth Tenant-to-tenant Migration and Copilot Takeoff
Strategize a Smooth Tenant-to-tenant Migration and Copilot TakeoffStrategize a Smooth Tenant-to-tenant Migration and Copilot Takeoff
Strategize a Smooth Tenant-to-tenant Migration and Copilot Takeoff
 
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...
 
Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...
 
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
 
Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
 
Mcleodganj Call Girls 🥰 8617370543 Service Offer VIP Hot Model
Mcleodganj Call Girls 🥰 8617370543 Service Offer VIP Hot ModelMcleodganj Call Girls 🥰 8617370543 Service Offer VIP Hot Model
Mcleodganj Call Girls 🥰 8617370543 Service Offer VIP Hot Model
 
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
 
Elevate Developer Efficiency & build GenAI Application with Amazon Q​
Elevate Developer Efficiency & build GenAI Application with Amazon Q​Elevate Developer Efficiency & build GenAI Application with Amazon Q​
Elevate Developer Efficiency & build GenAI Application with Amazon Q​
 
Understanding the FAA Part 107 License ..
Understanding the FAA Part 107 License ..Understanding the FAA Part 107 License ..
Understanding the FAA Part 107 License ..
 
MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024
 
Apidays New York 2024 - APIs in 2030: The Risk of Technological Sleepwalk by ...
Apidays New York 2024 - APIs in 2030: The Risk of Technological Sleepwalk by ...Apidays New York 2024 - APIs in 2030: The Risk of Technological Sleepwalk by ...
Apidays New York 2024 - APIs in 2030: The Risk of Technological Sleepwalk by ...
 
Navigating the Deluge_ Dubai Floods and the Resilience of Dubai International...
Navigating the Deluge_ Dubai Floods and the Resilience of Dubai International...Navigating the Deluge_ Dubai Floods and the Resilience of Dubai International...
Navigating the Deluge_ Dubai Floods and the Resilience of Dubai International...
 
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...
 
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
 
Architecting Cloud Native Applications
Architecting Cloud Native ApplicationsArchitecting Cloud Native Applications
Architecting Cloud Native Applications
 
Repurposing LNG terminals for Hydrogen Ammonia: Feasibility and Cost Saving
Repurposing LNG terminals for Hydrogen Ammonia: Feasibility and Cost SavingRepurposing LNG terminals for Hydrogen Ammonia: Feasibility and Cost Saving
Repurposing LNG terminals for Hydrogen Ammonia: Feasibility and Cost Saving
 
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
 

Normalized averaging using adaptive applicability functions with applications in image reconstruction from sparsely and randomly sampled data

  • 1. Normalized averaging using adaptive applicability functions with applications in image reconstruction from sparsely and randomly sampled data Presented at SCIA 2003 Tuan Q. Pham and Lucas J. van Vliet July 17, 2009 1 Pattern Recognition Group
  • 2. Overview • Normalized averaging • Local structure adaptive filtering • Experimental results • Comparison with diffusion-based image inpainting • Directions for further research July 17, 2009 2
  • 3. Normalized averaging (Knutsson’93) •Weighted average filtering: r = s*a •Normalized averaging = weighted average + signal/certainty principle: •each signal s is associated with a certainty c •s & c have to be processed separately ( s . c) ∗ a r= c∗a where s :signal, c :certainty, a :filter, r :result, * :convolution input with 10% Gaussian smoothing NA with Gaussian original pixels (σ = 1) applicability (σ = 1) July 17, 2009 3
  • 4. Normalized averaging: An example Reconstruction from 10% random pixels Nearest neighbor interpolation NA with adaptive applicability July 17, 2009 4
  • 5. Image reconstruction using Adaptive Normalized Averaging Input Image σ=1 Normalized Output Image (sparsely & Averaging (with local randomly structure sampled) Adaptive extended into applicability missing regions) Structure Analysis July 17, 2009 5
  • 6. Local structure adaptive filtering •Local structure from the structure tensor r v r r r T u ϕ rr T rr T T = ∇I ∇I = λu uu + λv vv r •orientation φ = arg(u) y = 1 κ x2 •anisotropy A = (λu - λv)/(λu + λv) 2 r •curvature κ = ∂φ /∂ v •scale rdensity = sample density •Scale-adaptive curvature-bent anisotropic Gaussian kernel with scales in 2 orthogonal directions: σ u = C (1 − A)α rdensity σ v = C (1 + A)α rdensity where C ~ SNR α ~ degree of structure enhancement kernel aligns with local structure July 17, 2009 6
  • 7. Sample Density Transform •Definition: Smallest radius of a pillbox, centered at each pixel, that encompasses total certainty of at least 1 •Role: Automatic scale selection of the applicability in the NA equation to avoid unnecessary smoothing Lena with missing hole Density transform NA with Gaussian(σ=1) Adap. Norm. Avg. July 17, 2009 7
  • 8. 4x4 super-resolution from 4 noisy frames • 4 input LowRes captured with fill-factor = 25%, intensity noise (σ=10), registration noise (σ=0.2 LR pitch) 1 of 4 input 64x64 LR SR using triangulation SR using adaptive NA • 16 times upsampling from only 4 frames. How is it possible: along linear structures, only 4 samples are enough for 4x super-resolution July 17, 2009 8
  • 9. Orientation Anisotropy Curvature Sample density Scale along linear structures Scale in perpendicular direction July 17, 2009 9
  • 10. Comparison with image inpainting • Image inpainting (Sapiro) = diffusion with level line evolution • also extending orientation into the missing regions • slow due to iterative nature • poor result for large holes input inpainting inpainting + Adapt. Norm. Avg. 110 iters (6 min) texture synthesis 0 iters (6 sec) July 17, 2009 10
  • 11. Directions for Further Research • Applications • Image filtering (noise/watermark removal, edge enhancement...) • Image interpolation from sparsely and randomly sampled data (image inpainting, image fusion, super-resolution...) • Further improvements • Scale-space local structure analysis. • Detect multiple orientations using orientation space. • Robust neighborhood operator than the weighted mean. July 17, 2009 11
  • 12. Image inpainting of thin scribbles input inpainting Adaptive Normalized Averaging (10 sec) July 17, 2009 12
  • 13. Simultaneous geometry/texture inpainting texture input geometry Adaptive NA (1 min) July 17, 2009 13
  • 14. Inpainting of ambiguous discontinuity original input inpainting Adaptive NA (1 sec) July 17, 2009 14