Motivation

1
Introduction
Motivation
Demonstration
Key Idea
1. Part detection

• Symmetric shape decompositions offer an effective
  representation for shape indexing and matching.
• Skeleton-based approaches assume correct
  figure-ground segmentation, precluding their
  application to cluttered scenes.
• Filter-based approaches are imprecise and offer
  poor precision-recall in both part detection and
  part grouping.
• Contour-based methods can suffer from high
  computational complexity and typically stop short
  of part grouping.
• How can we extract and group the symmetric parts
  of an object from a cluttered scene in a
  computationally efficient manner?

• Traditional skeletonization algorithms decompose
  a closed contour into symmetric parts and
  attachment relations a highly top-down process.
• In contrast, we will detect symmetric parts
  locally and then assemble them into an object a
  highly bottom-up process.
• We begin by segmenting the image into compact
  regions at multiple scales, representing a large
  set of (deformable) maximal disc hypotheses from
  which skeletal parts can be detected.
• Adjacent maximal disc hypotheses that form a
  symmetric part with strong image evidence are
  grouped.
• Finally, detected parts whose attachments are
  deemed nonaccidental are assembled to form
  objects.

1c. Group superpixels into symmetric parts using
the affinities
1b. Superpixel affinity (for each scale)
1a. Multiscale superpixel segmentation
Input image
Region skeleton
Blobs/Ridges (Lindeberg and Bretzner)
Contour Groups (Stahl and Wang)
2. Part grouping
Input image
2a. Part affinity
2b. Group parts
1. Part Detection
2. Part Grouping
Results
Approach
Object part
Maximal discs
Superpixel approximation
Approach

• Trained part detection and grouping on images
  from Weizmann Horse dataset (Borenstein and
  Ullman 2002).
• Quantitative evaluation of part detection on
  images from the Weizmann Horse dataset (compare
  to Lindeberg and Bretzner, 2003).
• Qualitative evaluation on images from other
  domains.

• A symmetric part is the locus of maximal disc
  centers (Blum 1967).
• Use superpixels at multiple scales as
  data-driven hypotheses of maximal discs.
• Identify groups of superpixels that havestrong
  symmetric edge support.

• Goal Group together parts likely to belong to
  the same object, based on detecting part
  attachments.
• Issue Since the same part can be detected at
  different scales, resulting groups may contain
  redundant parts.
• Solution
• Greedily cluster parts based on computed part
  attachment affinities.
• Decide which redundant parts to remove in the
  context of each cluster.

2a. Part affinity
1a. Multiscale superpixel segmentation
Too coarse
Too fine
Good scale

• Part affinity defined differently for redundant
  and attached parts.
• Redundant parts are assigned high affinity, while
  non-redundant parts are assigned an affinity
  based on evidence of attachment.
• Learn redundancy and attachment affinity from
  labeled training data.

• Parts appear at different scales in the image
• Need at least one superpixel scale that captures
  each part well
• Solution Use multiscale superpixel segmentation

image
main part clusters
image
main part clusters
Probability part not redundant
Redundant parts affinity
Probability that part is redundant
Attached parts affinity



1b. Superpixel affinity

• Define an adjacency graph for each superpixel
  scale.
• Each edge is assigned an affinity representing
  the likelihood that the two superpixels represent
  maximal discs of the same symmetric part.
• Affinity has shape and appearance components
• Shape check the presence of symmetric edges
• Appearance homogeneity
• Shape features
• Learn the affinity function from training data.

Part affinity

• Features to compute part attachment affinity
• Boundaryevidence
• Appearance similarity
• Attachment category (based on main part axes)

• Features to determine part redundancy
• Area overlap
• Boundary overlap
• Appearance similarity

Shape features
Appearance features
SVM
Logistic
Superpixel affinity Logistic
2b. Grouping parts into objects

• Cluster parts using affinities.
• Use standard graph clustering algorithm.(We use
  Felzenszwalbs and Huttenlochers greedy
  clustering algorithm (2004) )
• Select non-redundant parts from each cluster.
• Formulate as a quadratic program that minimizes
  overlap, of selected parts, while maximizing
  the covered area.

1c. Grouping superpixels into parts

• Cluster superpixels using affinities.
• Use standard graph clustering algorithm.(We use
  Felzenszwalbs and Huttenlochers greedy
  clustering algorithm (2004) )