Improved Subspace Clustering via Exploitation of Spatial Constraints

We present a novel approach to improving subspace clustering by exploiting the spatial constraints. The new method encourages the sparse solution to be consistent with the spatial geometry of the tracked points, by embedding weights into the sparse formulation. By doing so, we are able to correct sp...

Full description

Bibliographic Details
Main Authors: Pham, DucSon, Budhaditya, Saha, Phung, Dinh, Venkatesh, Svetha
Other Authors: N/A
Format: Conference Paper
Published: IEEE 2012
Subjects:
Online Access:http://hdl.handle.net/20.500.11937/4473
_version_ 1848744525600653312
author Pham, DucSon
Budhaditya, Saha
Phung, Dinh
Venkatesh, Svetha
author2 N/A
author_facet N/A
Pham, DucSon
Budhaditya, Saha
Phung, Dinh
Venkatesh, Svetha
author_sort Pham, DucSon
building Curtin Institutional Repository
collection Online Access
description We present a novel approach to improving subspace clustering by exploiting the spatial constraints. The new method encourages the sparse solution to be consistent with the spatial geometry of the tracked points, by embedding weights into the sparse formulation. By doing so, we are able to correct sparse representations in a principled manner without introducing much additional computational cost. We discuss alternative ways to treat the missing and corrupted data using the latest theory in robust lasso regression and suggest numerical algorithms so solve the proposed formulation. The experiments on the benchmark Johns Hopkins 155 dataset demonstrate that exploiting spatial constraints significantly improves motion segmentation.
first_indexed 2025-11-14T06:02:51Z
format Conference Paper
id curtin-20.500.11937-4473
institution Curtin University Malaysia
institution_category Local University
last_indexed 2025-11-14T06:02:51Z
publishDate 2012
publisher IEEE
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-44732017-09-13T16:02:18Z Improved Subspace Clustering via Exploitation of Spatial Constraints Pham, DucSon Budhaditya, Saha Phung, Dinh Venkatesh, Svetha N/A subspace segmentation sparse representation clustering We present a novel approach to improving subspace clustering by exploiting the spatial constraints. The new method encourages the sparse solution to be consistent with the spatial geometry of the tracked points, by embedding weights into the sparse formulation. By doing so, we are able to correct sparse representations in a principled manner without introducing much additional computational cost. We discuss alternative ways to treat the missing and corrupted data using the latest theory in robust lasso regression and suggest numerical algorithms so solve the proposed formulation. The experiments on the benchmark Johns Hopkins 155 dataset demonstrate that exploiting spatial constraints significantly improves motion segmentation. 2012 Conference Paper http://hdl.handle.net/20.500.11937/4473 10.1109/CVPR.2012.6247720 IEEE restricted
spellingShingle subspace segmentation
sparse representation
clustering
Pham, DucSon
Budhaditya, Saha
Phung, Dinh
Venkatesh, Svetha
Improved Subspace Clustering via Exploitation of Spatial Constraints
title Improved Subspace Clustering via Exploitation of Spatial Constraints
title_full Improved Subspace Clustering via Exploitation of Spatial Constraints
title_fullStr Improved Subspace Clustering via Exploitation of Spatial Constraints
title_full_unstemmed Improved Subspace Clustering via Exploitation of Spatial Constraints
title_short Improved Subspace Clustering via Exploitation of Spatial Constraints
title_sort improved subspace clustering via exploitation of spatial constraints
topic subspace segmentation
sparse representation
clustering
url http://hdl.handle.net/20.500.11937/4473