Title: Skeleton-search: Category-specific object segmentation/recognition using a skeletal shape model
1Skeleton-search Category-specific object
segmentation/recognition using a skeletal shape
model
- Nhon Trinh Benjamin Kimia
- Brown University
- British Machine Vision Conference
- Sep 08, 2009
2Category-specific object recognition
- Q Is there a giraffe in this image?
- A Yes.
- Q Really? Can you delineate it?
3(No Transcript)
4(No Transcript)
5(No Transcript)
6(No Transcript)
7Faces of Object Recognition
increasing difficulty
Segmentation Part labeling
Object detection
Object segmentation
Image classification
8Our goal
9Top-down approach
Space of Giraffe Shapes
- How to represent shapes of an object category?
- How to measure support for a shape in an image?
- How to search for the best supported shapes?
10Contributions
- How to represent shapes of an object category?
- ? Fragment-Based Generative Model for Shape
- How to measure support for a shape in an image?
- ? Improvement to Oriented Chamfer Matching
- How to search for the best supported shapes?
- ? Extension to the Viterbi algorithm to compute
multiple solutions
111. How to represent shapes of an object category?
12Giraffes in Images
Courtesy of Vittorio Ferrari
13Giraffe Shapes
14Giraffe Skeleton
15Shared Skeletal Topology
16Idea Represent a shape using its skeleton
17Intrinsic Symmetry-based Shape Model (Trinh and
Kimia (ICCV07)
18Intrinsic Shape Model for Segmentation
- Drawback global dependency of each fragments
boundary on other fragments. - New model able to reconstruct each fragment
LOCALLY from its adjacent nodes.
19Fragment-Based Generative Model for Shape
Parameter Set
20Reconstructing a Shape Fragments Boundary
- Interpolate A?B and D?C contours using smooth
bi-arcs (Kimia et al., IJCV 2003).
21Generative Model
222. How to measure support for a shape in an image?
23Cost function
- Cost of a shape sum of its fragments costs.
24Cost of a shape fragment
shape prior
image support
cost of fragment
- Shape prior uniform distribution on the
fragments intrinsic parameters. - Image support
- Region appearance
- Edge support for pair of boundary contours
25Oriented Chamfer Matching (OCM)(Shotton et al,
PAMI08 and Jain et al, CVIU07)
26Oriented Chamfer Matching (OCM)(Shotton et al,
PAMI08 and Jain et al, CVIU07)
- Match each contour point to its closest edge
27Drawbacks of OCM
- Over-counting support when edges missing.
- Under-counting support when many spurious edges
present. - Awarding accidental alignment.
28Improvement Contour Chamfer Matching (CCM)
- Partition edges into thin stripes.
- Match contour points to image edges using OCM
cost. - Penalize orientation discrepancies between query
contour and the contour connecting image edges.
29How to search for the best supported shapes?
30Single Global Solution
- Use Viterbi algorithm on a tree.
31The need for multiple solutions
32Single-Pass Multiple Solution Using DP
- Candidate pool optimal solutions for each
position of root node. - Differential Exclusion Principle
- Trimming discarding non-max solutions the
candidate pool.
33Experiments
34Dataset ETHZ Shape Classes
- 255 images
- 5 categories giraffes, bottles, applelogos,
swans, mugs.
Courtesy of Vittorio Ferrari
35Detection / Segmentation - Giraffes
model
36Detection/Segmentation - Bottles
37Detection/Segmentation - Swans
38Detection/Segmentation - Applelogos
Model
39Detection/Segmentation - Mugs
40False Positives
41Object Detection Evaluation
ground-truth
detection
42Object Detection Performance
43Evaluation Segmentation Performance
- (Ferrari et al, INRIA Tech Report 2008)
- Boundary Coverage proportion of the ground-truth
that is close to the segmented shape. - Boundary Precision proportion of the segmented
shape that is close to the ground-truth.
Boundary precision
Boundary coverage
segmentation output
44Performance boundary coverage
45Performance Boundary Precision
46Summary
- A skeleton-based generative model for shape where
each fragment can be reconstructed locally. - Improvement to Oriented Chamfer Matching cost.
- Extension to Viterbi algorithm to compute
multiple solutions in a single pass.
47Thank youQuestions?
- Email ntrinh_at_lems.brown.edu