Your SlideShare is downloading. ×
0
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Lexie.IGARSS11.pptx
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×
Saving this for later? Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime – even offline.
Text the download link to your phone
Standard text messaging rates apply

Lexie.IGARSS11.pptx

137

Published on

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
137
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
8
Comments
0
Likes
0
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide
  • The added earth logo is from the website: http://rst.gsfc.nasa.gov/Sect19/Sect19_2a.html
  • PREVIOUS WORK TO SOLVE THE DIFFICULTIES: Semi-supervised approach requires the assumption of smooth changes. However, sometimes the assumption maybe not true for multitemporal data sets.It is also commonly seen that adaptive schemes are used to redefine decision boundaries. Statistically speaking, class distributions will alter due to environments. Mean or variances will be different from a scene to a scene. Decision boundaries therefore are needed to adjust according to samples from new scene. DIFFERENT POINT OF VIEW: in geometric learning point of view, since we are talking about a geometric learning methodology, we assume two data sets are similar in some sense, and we need to find a mapping between two similar structures.
  • WHY DOES MA WORK FOR CLASSIFICATION: Our main interest is to classify. Aligning similar underlying manifolds is beneficial to classification work when at least one image contains label information. A joint manifold can characterize geometric structures of both data sets.
  • First term: preserving given featuresSecond term: clustering conditions on local properties\\mu: tuning the relative weights of two terms in the cost function
  • First term: pairwise alignment constraintsSecond and third terms: Local properties\\mu: tuning the relative weights of two terms in the cost function
  • Font in equation description
  • BASELINE: Demonstrate how pooled data can fail a proper joint manifoldUse other colors, not gray
  • Change color MA space using lower cases
  • Bold: class accuracy May, June pair
  • [WK] Slide 5 introduced the abbreviations for the local methods.
  • [WK] Slide 5 introduced the abbreviations for the local methods.
  • Compare to previous results?
  • Transcript

    • 1. Manifold Alignment for MultitemporalHyperspectral Image Classification
      H. Lexie Yang1, Dr. Melba M. Crawford2
      School of Civil Engineering, Purdue University
      and
      Laboratory for Applications of Remote Sensing
      Email: {hhyang1, mcrawford2}@purdue.edu
      July 29, 2011
      IEEE International Geoscience and Remote Sensing Symposium
    • 2. Outline
      Introduction
      Research Motivation
      Effective exploitation of information for multitemporal classification in nonstationary environments
      Goal: Learn “representative” data manifold
      Proposed Approach
      Manifold alignment via given features
      Manifold alignment via correspondences
      Manifold alignment with spectral and spatial information
      Experimental Results
      Summary and Future Directions
    • 3. Introduction
      N>>30
      3
      2
      1
      2001
      2003
      2004
      2005
      2006
      2002
      2001
      N narrow spectral bands
      June
      July
      May
      May
      May
      May
      June
      Challenges for classification of hyperspectral data
      temporally nonstationary spectra
      high dimensionality
    • 4. Research Motivation
      Nonstationarities in sequence of images
      Spectra of same class
      may evolve or drift
      over time
      Potential approaches
      Semi-supervised methods
      Adaptive schemes
      Exploit similar data geometries
      Explore data manifolds
      Good initial conditions required
    • 5. Manifold Learning for Hyperspectral Data
      Characterize data geometry with manifold learning
      To capture nonlinear structures
      To recover intrinsic space (preserve spectral neighbors)
      To reduce data dimensionality
      Classification performed in low dimensional space
      Original space
      Manifold space
      3rd dim
      Spectral bands
      n
      Spatial dimension
      6
      5
      4
      3
      2
      1
      Spatial dimension
      1st dim
      2nd dim
    • 6. Challenges: Modeling Multitemporal Data
      • Unfaithful joint manifold
      due to spectra shift
      • Often difficult to model the inter-image correspondences
      Data manifold at T2
      Data manifolds at T1 and T2
      Data manifold at T1
    • 7. Proposed Approach: Exploit Local Structure
      • Assumption: local geometric structures are similar
      • 8. Approach: Extract and optimally align local geometry to minimize overall differences
      Locality
      Spectral space at T2
      Spectral space at T1
    • 9. Proposed Approach: Conceptual Idea
      (Ham, 2005)
    • 10. Proposed Approach: Manifold Alignment
      • Exploit labeled data for classification of multitemporal data sets
      Samples with class labels
      Samples with no class labels
      Joint manifold
    • 11. Manifold Alignment: Introduction
      and are 2 multitemporalhyperspectral images
      Predict labels of using labeled
      Explore local geometries using graph Laplacian and some form of prior information
      Define Graph Laplacian
      Twopotential forms of prior information: given features and pairwise correspondences [Ham et al. 2005]
    • 12. Manifold Alignment via Given Features
      Minimize
      Joint Manifold
      Given Features
    • 13. Manifold Alignment via Pairwise Correspondences
      Minimize
      Correspondences between and
      Joint Manifold
    • 14. MA with spectral and spatial information
      Combine spatial locations with spectral signatures
      To improve local geometries (spectral) quality
      Idea: Increase similarity measure when two samples are close together
      Weight matrix for graph Laplacian:
      where spatial location of each pixel is represented as
    • 15. Experimental Results: Data
      • Three Hyperion images collected in May, June and July 2001
      • 16. May, June pair: Adjacent geographical area
      • 17. June, July pair: Targeted the same area
      May June July
    • 18. Experimental Results: Framework
      L
      L
      L
      I1, I2
      I1
      I2
      Graph Laplacian
      Prior information
      Joint manifold
      Given features
      Classification
      with KNN
      Correspondences
      Develop Data Manifold of Pooled Data
    • 19. Manifold Learning for Feature Extraction
      Global methods consider geodesic distance
      Isometric feature mapping (ISOMAP)
      Local methods consider pairwise Euclidian distance
      Locally Linear Embedding (LLE): (Saul and Roweis, 2000)
      Local Tangent Space Alignment (LTSA): (Zhang and Zha, 2004)
      LaplacianEigenmaps (LE): (Belkin and Niyogi, 2004)
      (Tenenbaum, 2000)
    • 20. MA with Given Features
      Baseline: Joint manifold developed by pooled data
      79.21
      77.29
      77.88
      76.31
      (May, June pair)
    • 21. MA Results – Classification Accuracy
      • Evaluate results by overall accuracies
    • Results – Class Accuracy
      May, June pair
      Typical class
      Critical class
      (Island Interior)
      Critical class
      (Riparian)
      (Woodlands)
    • 22. Summary and Future Directions
      Multitemporal spectral changes result in failure to provide a faithful data manifold
      Manifold alignment framework demonstrates potential for nonstationary environment by utilizing similar local geometries and prior information
      Spatial proximity contributes to stabilization of local geometries for manifold alignment approaches
      Future directions
      Investigate alternative spatial and spectral integration strategy
      Address issue of longer sequences of images
    • 23. Thank you.
      Questions?
    • 24. References
      J. Ham, D. D. Lee, and L. K. Saul, “Semisupervised alignment of manifolds,” in International Workshop on Artificial Intelligence and Statistics, August 2005.
    • 25. Backup Slides
    • 26. Local Manifold Learning for Feature Extraction (s,f)
      Local geometry preserved via various strategies for embedding
      Popular local manifold learning methods
      Locally Linear Embedding (LLE): (Saul and Roweis, 2000)
      Local Tangent Space Alignment (LTSA): (Zhang and Zha, 2004)
      LaplacianEigenmaps (LE): (Belkin and Niyogi, 2004)
      Pairwise distance between neighbors computed using Gaussian kernel function - O(pN2) method
      Embedding computed to minimize the total distance between neighbors
    • 27. LE: Impact of Parameter Values
      Parameter values for local embedding
      s obtained via grid search
      k, p obtained empirically
      BOT Class 3, 6
      BOT Classes 1-9
    • 28.
      • Island Interior
      Alignment Results: Typical Class
    • 29.
      • Critical class: Riparian
      Alignment Results: Critical Class
    • 30. Alignment Results: Critical Class
      Critical class: Woodlands
    • 31. MA Results – Classification Accuracy
      • Evaluate results by overall accuracies
      Labeled Class
      (Subset Data)
      Classified via
      Given Features
      Classified via Correspondences
      Classified via
      Pooled Data
      May, June pair
    • 32. MA Results – Classification Accuracy
      Classified via
      Given Features
      (Spectral + spatial)
      Classified via Correspondences
      (Spectral + spatial)
      Labeled Class
      (Subset Data)
      Classified via
      Given Features
      (Spectral)
      Classified via Correspondences
      (Spectral)
      May, June pair

    ×