for Object Segmentation
University of Texas at Austin, Jaechul Kim and Kristen Grauman
(a) Sematically close
Intuition for shape sharing. While one may expect shape sharing between
objects of semantically close categories (a), we observe that similar
even among semantically disparate objects (b). This
suggests to transfer “object-level” shapes between categories, enabling
category-independent shape priors for object segmentation.
observe shapes are shared among different classes and apply this
phenomenon to category-independent object segmentation by trasnferring
a shape from one class to (possibly unknown) another class.
idenitfy and transfer the shared shapes between images, we propose a
non-parametric local shape matching using our BPLR local shape
- Not only do we show the shape sharing enhances the
segmentation quality of unseen, novel objects on the standard benchmark dataset, but we also
provide a deeper analysis on when and why sharing is really useful in
terms of color easiness and object size.
Basis of Approach: Shape
Trasnfer by Partial Shape Matching
Basis of approach: Transfer category-indepedent shape prior through
partial shape matching.
We see a shape
of a bird in the exemplar is transfered to a person of the test image,
where we did not assume any category specific knolwedges on those
Overview of the Method
Figure 3. Overview of our method. (a) Exemplars (first column) that
share shape with the test image (second column) are projected in, no
matter their category.
identify those shared shapes (marked in blue) via local BPLR matches
(marked in white). (b) Multiple exemplars that highly overlap are
aggregated, to form
a shape prior and color model per each
aggregated group. (c) The priors from each group are used to
compute a series of graph-cut segmentation hypotheses.
Segmentation accuracy in benchmark datasets
* Note: For both experiments, we use the common PASCAL training dataset for exemplars.
PASCAL 2010 validation dataset
(b) Berkeley Segmentation Dataset
When does shape sharing help most?
We evaluate gain by shape sharing as a function of color easiness and object size.
Color easinesss basically means how easily (or accurately) we can
segment a given image using only its color information (see the paper
4. Shape Sharing's gain as a function of color easiness and
object size. The gain is compared to CPMC [Carreira and Sminchisescu,
As color easinesss is getting lower, our gain
increases. In fact, at the leftmost point in the graph, where
color is hardest, we achieve as much as 15 points gain over the
Also, as object size is getting bigger, our gain
increases. This is intuitive because as object size becomes
larger, its shape becomes more visable.
Which classes share shapes?
Figure 5. Shape sharing matrix for the 20 classes in PASCAL.
compute the strength of sharing by counting how many times shape
exemplars from one class are used to generate the best segmentation
hypotheses for another class. Brighter=higher.
6. Example segmentations. Shape sharing is useful when objects consist
of diversely-colored parts (1st example). Also, it helps delineate
objects from confusing surrounding colors (2nd example).
shapes do not help segment tiny objects (3rd example), nor objects
lacking shape like truncated sofa in the 4th example.
Code + Data [zip]
 Shape Sharing for Object Segmentation. J. Kim and K. Grauman. ECCV 2012 [pdf
 Boundary Preserving Dense Local Regions. J. Kim and K. Grauman. CVPR 2011 [pdf
Last modified 11-20-2012