Unsupervised shape clustering using diffusion map
Rajpoot, Nasir M. (Nasir Mahmood) and Arif, Muhammad. (2008) Unsupervised shape clustering using diffusion map. Annals of the BMVA, Volume 2008 (Number 5).
WRAP_Rajpoot_2008-0005.pdf - Published Version - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (738Kb) | Preview
The quotient space of all smooth and connected curves represented by a ﬁxed number of boundary points is a ﬁnite-dimensional Riemannian manifold, also known as a shape manifold. This makes the preservation of locality a critically important issue when reducing the dimensionality of shapes on the manifold. We present a completely unsupervised clustering algorithm employing diffusion maps for locality-preserving embedding of shapes onto a much lower-dimensional space. The algorithm ﬁrst obtains a non-linear low-dimensional embedding of shape context features of outer boundary contours of the shapes. Considering the embedded coordinates as a new minimalist representation of shapes, a clustering of shapes is obtained using a ﬁnite mixture model. The proposed clustering algorithm is computationally efﬁcient, as it relies on clustering in a very lowdimensional space, and produces much improved results (88.6% for a 7-class dataset) as compared to clustering with conventional linear projections.
|Item Type:||Journal Article|
|Subjects:||Q Science > QA Mathematics
Q Science > QA Mathematics > QA76 Electronic computers. Computer science. Computer software
|Divisions:||Faculty of Science > Computer Science|
|Library of Congress Subject Headings (LCSH):||Riemannian manifolds, Algorithms|
|Journal or Publication Title:||Annals of the BMVA|
|Publisher:||The British Machine Vision Association and Society for Pattern Recognition|
|Number of Pages:||17|
|Access rights to Published version:||Restricted or Subscription Access|
M. Belkin and P. Niyogi. Laplacian eigenmaps for dimensionality reduction and data representation.
Actions (login required)