Unsupervised Category Modeling, Recognition and Segmentation

1
Unsupervised Category Modeling, Recognition and
Segmentation

• Sinisa Todorovic and Narendra Ahuja

2
What is Common in a Set of Images?
Images possibly contain an object of interest
Which objects appear frequently in the set?
What properties are shared by similar objects in
the set?
Where are the objects?
3
Objective Car Example
occlusion
no car
multiple cars
occlusion
4
Problem Definition
GIVEN
Images possibly containing frequent occurrences
of similar objects
DETERMINE
Training
If similar objects are present
AND IF YES LEARN
The model of similar objects
5
Prior Work Dominated By

• Statistical modeling of local features patches
  or curve fragments
• Trend Object detection Image classification
• Trend Object segmentation Object localization
• Trend Object segmentation Binary thresholding
  of a probabilistic map
• Hypothesize the number of objects and their parts
• Hypothesize the topology of object parts
• Each training image must contain a category of
  interest
• Modeling background
• Require typically hundreds of training images

6
Category Modeling is Very Difficult

• Explicit modeling of recursive embedding of
  object subparts
• Regions vs. local features open questions
• More informative?
• More stable and robust to noise?
• Regions allow
• simultaneous object detection and segmentation
• explicit representation of the recursive
  embedding property

7
Our Approach
SIMILAR OBJECTS PRESENT IN THE SET
MANY SUBIMAGES WITH SIMILAR REGION PROPERTIES
ABUNDANT DATA
ROBUST LEARNING IS FEASIBLE
8
Region Properties

• Geometric
• Region area
• Boundary shape
• Photometric
• Gray-level contrast with the surround
• Topology
• Recursive containment of regions
• Layout - relative region locations

9
Feature Extraction Image Segmentation
segmentation
Homogeneous regions at ALL contrasts and sizes
Image
N. Ahuja TPAMI 96, Tabb Ahuja TIP 97, Arora
Ahuja ICPR 06
10
Multiscale Segmentation to Segmentation Tree
Sample cutsets
Segmentation tree
Example segmentations
Contrast level ? Tree level
11
Image Tree and Object Subtree
12
Outline of Our Approach
Images Trees
Category present Many similar subtrees
Extracting similar subtrees Tree matching
Category model Union of similar subtrees
Simultaneous detection, recognition and
segmentation of ALL category instances by Matching
the model with an image
13
Tree Matching Structural Noise
Edit-distance tree matching
Pelillo et a. PAMI99, TorselloHancock ECCV02,
PRL03
14
Matching Algorithm
Input trees
Matched subtrees
15
Matching Algorithm
GIVEN two trees
FIND bijection
which MAXIMIZES their similarity measure
node saliency
cost of node matching
while PRESERVING ancestor-descendant relationships
16
Matching Algorithm Recursive Solution
descendants
Maximum clique over all descendant pairs
17
Outline
LEARNING
18
Category Model Tree Union
Tree intersection
Tree union
19
Simultaneous Detection and Segmentation
MATCHING
20
Performance Evaluation Criteria
DETECTION ERROR
Matched Subtrees (MST)
Ground Truth (GT)
False positive intersection of MST and GT lt
0.5 union of MST and GT
SEGMENTATION ERROR
Matched Subtrees (MST)
Ground Truth (GT)
XOR of MST and GT
21
Results UIUC Cars Side View
Results on test images
22
Results Faces -- Caltech 101 Database
Results on test images
23
Results Caltech Cars Rear View
10 positive out of 20 training images
24
Recall-Precision
Training from a small-size dataset
Varying tradeoff recall vs. precision
25
Complexity and Runtime on 2.4GHZ 2GB RAM PC
of tree nodes
Training on 20 images of UIUC CARS lt 2 hours
26
Summary and Conclusion

• Unsupervised learning of an unknown category
  frequently occurring in a given set of images
• Region-based, structural approach
• Simultaneous detection, recognition, and
  segmentation of all category instances in unseen
  images
• NO multiple detections on the same object
• NO hypotheses on the number of objects and their
  parts
• NO hypotheses on the topology of object parts
• Small number of training images
• Complexity comparable with standard methods

27
Acknowledgment
THANK YOU!
sintod, n-ahuja_at_uiuc.edu
http//vision.ai.uiuc.edu