Http:// Berkeley Vision GroupNIPS Vancouver 20021 Learning to Detect Natural Image Boundaries Using...
-
date post
20-Dec-2015 -
Category
Documents
-
view
214 -
download
0
Transcript of Http:// Berkeley Vision GroupNIPS Vancouver 20021 Learning to Detect Natural Image Boundaries Using...
NIPS Vancouver 2002 1http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Learning to Detect Natural Image Boundaries Using Local Brightness,
Color, and Texture Cues
David Martin, Charless Fowlkes, Jitendra Malik{dmartin,fowlkes,malik}@eecs.berkeley.edu
UC Berkeley Vision Grouphttp://www.cs.berkeley.edu/projects/vision
NIPS Vancouver 2002 3http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Multiple Cues for Grouping
• Many cues for perceptual grouping: – Low-Level: brightness, color, texture, depth, motion
– Mid-Level: continuity, closure, convexity, symmetry, …
– High-Level: familiar objects and configurations
This talk: Learn local cue combination rule from data
NIPS Vancouver 2002 4http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Non-Boundaries Boundaries
I
T
B
C
NIPS Vancouver 2002 5http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Goal and Outline
• Goal: Model the posterior probability of a boundary Pb(x,y,) at each pixel and orientation using local cues.
• Method: Supervised learning using dataset of 12,000 segmentations of 1,000 images by 30 subjects.
• Outline of Talk:
1. 3 cues: brightness, color, texture
2. Cue calibration
3. Cue combination
4. Compare with other approaches– Canny 1986, Konishi/Yuille/Coughlan/Zhu 1999
5. Pb images
NIPS Vancouver 2002 6http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Brightness and Color Features
• 1976 CIE L*a*b* colorspace• Brightness Gradient BG(x,y,r,)
2 difference in L* distribution
• Color Gradient CG(x,y,r,) 2 difference in a* and b*
distributions
i ii
ii
hg
hghg
22 )(
2
1),(
r(x,y)
NIPS Vancouver 2002 7http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Texture Feature
• Texture Gradient TG(x,y,r,) 2 difference of texton histograms
– Textons are vector-quantized filter outputs
TextonMap
NIPS Vancouver 2002 8http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Cue Calibration
All free parameters optimized on training data
• Brightness Gradient– Scale, bin/kernel sizes for KDE
• Color Gradient– Scale, bin/kernel sizes for KDE, joint vs. marginals
• Texture Gradient– Filter bank: scale, multiscale? – Histogram comparison: L1, L2, L, 2, EMD– Number of textons– Image-specific vs. universal textons
• Localization parameters for each cue (see paper)
NIPS Vancouver 2002 9http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Classifiers for Cue Combination
• Classification Trees– Top-down splits to maximize entropy, error bounded
• Density Estimation– Adaptive bins using k-means
• Logistic Regression, 3 variants– Linear and quadratic terms– Confidence-rated generalization of AdaBoost (Schapire&Singer)
• Hierarchical Mixtures of Experts (Jordan&Jacobs)– Up to 8 experts, initialized top-down, fit with EM
• Support Vector Machines (libsvm, Chang&Lin)
Range over bias/variance, parametric/non-parametric, simple/complex
NIPS Vancouver 2002 10http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
ClassifierComparison
NIPS Vancouver 2002 11http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Cue Combinations
NIPS Vancouver 2002 12http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Alternate Approaches
• Canny Detector– Canny 1986
– MATLAB implementation
– With and without hysteresis
• Second Moment Matrix– Nitzberg/Mumford/Shiota 1993
– cf. Förstner and Harris corner detectors
– Used by Konishi et al. 1999 in learning framework
– Logistic model trained on eigenspectrum
NIPS Vancouver 2002 13http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Two Decades of Boundary
Detection
NIPS Vancouver 2002 14http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Pb Images ICanny 2MM Us HumanImage
NIPS Vancouver 2002 15http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Pb Images IICanny 2MM Us HumanImage
NIPS Vancouver 2002 16http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Pb Images IIICanny 2MM Us HumanImage
NIPS Vancouver 2002 17http://www.cs.berkeley.edu/projects/visionUC Berkeley Vision Group
Summary and Conclusion
1. A simple linear model is sufficient for cue combination– All cues weighted approximately equally in logistic
2. Proper texture edge model is not optional for complex natural images
– Texture suppression is not sufficient!
3. Significant improvement over state-of-the-art in boundary detection
– Pb useful for higher-level processing
4. Empirical approach critical for both cue calibration and cue combination
Segmentation data and Pb images on the webhttp://www.cs.berkeley.edu/projects/vision