geodesic minimal paths vida movahedi elder lab, january 2010
TRANSCRIPT
![Page 1: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/1.jpg)
Geodesic Minimal PathsGeodesic Minimal Paths
Vida MovahediVida Movahedi
Elder Lab, January 2010Elder Lab, January 2010
![Page 2: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/2.jpg)
ContentsContents
• What is the goal?
• Minimal Path Algorithm
• Challenges
• How can Elderlab help?
• Results
![Page 3: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/3.jpg)
GoalGoal
• Finding boundary of salient objects in images of natural scenes
![Page 4: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/4.jpg)
Minimal PathMinimal Path
• Inputs: – Two key points
– A potential function to be minimized along the path
• Output:– The minimal path
![Page 5: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/5.jpg)
Minimal Path- problem formulationMinimal Path- problem formulation
• Global minimum of the active contour energy:
C(s): curve, s: arclength, L: length of curve
• Surface of minimal action U: minimal energy integrated along a path between p0 and p
Ap0,p : set of all paths between p0 and p
],0[
))((~
)(L
dssCPCE
dssCPCEpUpoppop
ΑΑ)(
~inf)(inf)(
,,
![Page 6: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/6.jpg)
Fast Marching AlgorithmFast Marching Algorithm
• Computing U by frontpropagation: evolving a front starting from an infinitesimal circle around p0 until each point in image is reached
![Page 7: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/7.jpg)
ChallengesChallenges
• Can the minimal path algorithm solve the boundary detection problem?– Key points?
– Potential Function?
• Idea: Use York’s multi-scale algorithm (MS)
![Page 8: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/8.jpg)
MS AlgorithmMS Algorithm
• We have a set of contour hypotheses at each scale
• These contours can be used to find good candidates for key points
• These contours (and some other cues) can also be used to build potential functions.
• Multi-scale model (coarse to fine) can also help
![Page 9: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/9.jpg)
![Page 10: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/10.jpg)
Key PointsKey Points
• Simplest approach: 3 key points, equally spaced on the MS contour (prior)
• Maximize product of probabilities (MS unary cue)
![Page 11: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/11.jpg)
Rotating Key PointsRotating Key Points
• Consider multiple hypothesis for key points
• Obtain multiple contours
• Next step: Find which contour is the best– Distribution model for contour lengths
– Distribution model for average Pb value
– Improve method to find simple contours only
![Page 12: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/12.jpg)
Rotating Key PointsRotating Key Points
![Page 13: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/13.jpg)
Potential FunctionPotential Function
• Ideas:– The Sobel edge map
– Distance transform of MS contour (prior)
– Distance transform of several overlapped MS contours
– Berkeley’s Pb map
– Likelihood based on Pb and distance to prior contour
![Page 14: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/14.jpg)
Sobel Edge MapSobel Edge Map
![Page 15: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/15.jpg)
Sobel Edge MapSobel Edge Map
• Can use the MS prior to emphasize or de-emphasize map
![Page 16: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/16.jpg)
Distance TransformDistance Transform
![Page 17: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/17.jpg)
Distance transformDistance transform
• Too much emphasis on MS prior
![Page 18: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/18.jpg)
Distance transform Distance transform of 10 overlapped MS contoursof 10 overlapped MS contours
![Page 19: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/19.jpg)
Challenge: Challenge: If MS contours are not goodIf MS contours are not good
![Page 20: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/20.jpg)
Challenge: Challenge: If MS contours are not goodIf MS contours are not good
![Page 21: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/21.jpg)
Berkeley’s Pb mapBerkeley’s Pb map
![Page 22: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/22.jpg)
Combining Pb and DistanceCombining Pb and Distance
)|(
)|(
)|(
)|()()(),(
CxDp
CxDp
CxPbp
CxPbpDLPbLDPbL
Next step: learning models
![Page 23: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/23.jpg)
SummarySummary
• The MP algorithm provides global minimal paths
• The MS algorithm provides contour hypothesis
• The MS contours can be used to obtain key points and potential functions for MP algorithm
• Next steps:
– Learning models for better potential functions
– Obtaining simple contours
– Ranking contours
– Evaluate multi-scale model
![Page 24: Geodesic Minimal Paths Vida Movahedi Elder Lab, January 2010](https://reader036.vdocuments.us/reader036/viewer/2022062308/56649e5d5503460f94b569e9/html5/thumbnails/24.jpg)
ReferencesReferences
Laurent D. Cohen (2001), “Multiple Contour Finding and Perceptual Grouping using Minimal Paths”, Journal of Mathematical Imaging and Vision, vol. 14, pp. 225-236.
Estrada, F.J. and Elder, J.H. (2006) “Multi-scale contour extraction based on natural image statistics”, Proc. IEEE Workshop on Perceptual Organization in Computer Vision, pp. 134-141.
J. H. Elder, A. Krupnik and L. A. Johnston (2003), "Contour grouping with prior models," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, pp. 661-674.