arrow
Return

Task-Specific Image Partitioning

delete2013-02-01
delete16
delete
OA
AI
S
Sungwoong Kim *
S
Sebastian Nowozin
P
Pushmeet Kohli
C
Chang D. Yoo
DOI:10.1109/TIP.2012.2218822delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
Image partitioning is an important preprocessing step for many of the state-of-the-art algorithms used for performing high-level computer vision tasks. Typically, partitioning is conducted without regard to the task in hand. We propose a task-specific image partitioning framework to produce a region-based image representation that will lead to a higher task performance than that reached using any task-oblivious partitioning framework and existing supervised partitioning framework, albeit few in number. The proposed method partitions the image by means of correlation clustering, maximizing a linear discriminant function defined over a superpixel graph. The parameters of the discriminant function that define task-specific similarity/dissimilarity among superpixels are estimated based on structured support vector machine (S-SVM) using task-specific training data. The S-SVM learning leads to a better generalization ability while the construction of the superpixel graph used to define the discriminant function allows a rich set of features to be incorporated to improve discriminability and robustness. We evaluate the learned task-aware partitioning algorithms on three benchmark datasets. Results show that task-aware partitioning leads to better labeling performance than the partitioning computed by the state-of-the-art general-purpose and supervised partitioning algorithms. We believe that the task-specific image partitioning paradigm is widely applicable to improving performance in high-level image understanding tasks.
Keywords:
Correlation clustering
image partitioning
linear programming relaxation
structured support vector machine

Journal

IEEE Transactions on Image Processing cover
IEEE Transactions on Image Processing
IF:
13.7
Papers:
1.0W
Citations:
8.4W

Organization

Q
qualcomm
Scholars:
782
Papers: 646
Citations: 1
M
Microsoft
Scholars:
3.0K
Papers: 2.7K
Citations: 7
Cited Papers

Cited Papers

Avian community response to a golf-course landscape unit gradient
err2005-06-01
err0
PREAI
errStephen G. Jones; David H. Gordon; Gary M. Phillips; Bruce R. D. Richardson
errShare
errSave
The Randolph Glacier Inventory: a globally complete inventory of glaciers
err2017-07-10
err0
errOAAI
errW. Tad Pfeffer; Anthony A. Arendt; Andrew Bliss; Tobias Bolch; J. Graham Cogley; Alex S. Gardner; Jon-Ove Hagen; Regine Hock; Georg Kaser; Christian Kienholz; Evan S. Miles; Geir Moholdt; Nico Mölg; Frank Paul; Valentina Radić; Philipp Rastner; Bruce H. Raup; Justin Rich; Martin J. Sharp
errShare
errSave
Reported self-control is not meaningfully associated with inhibition-related executive function: A Bayesian analysis
err
IF0
err2017-08-17
err0
errOAAI
errBlair Saunders; Marina Milyavskaya; Alexander Etz; Daniel Randles; Michael Inzlicht
errShare
errSave
Synthesis and Chromatographic Properties of Liquid Crystalline Polysiloxanes Containing Steroid Substituents
err2007-03-28
err0
PREAI
errNathan W. Adams; Jerald S. Bradshaw; Jose-Maria Bayona; Karin E. Markides; Milton L. Lee
errShare
errSave
Experiencia, complicaciones y factores pronósticos con el reservorio anal en la colitis ulcerosa. Estudio observacional
err2020-02-01
err0
PREAI
errJavier Die; Juan Ocaña; Pedro Abadía; Juan Carlos García; Irene Moreno; Juan Diego Pina; Gloria Rodrígez; José Manuel Devesa
errShare
errSave
Lizard Abundance in an Exurban Southwestern Savanna, and the Possible Importance of Roadrunner Predation
err2006-04-01
err0
PREAI
errBLAKE W. AUDSLEY; CARL E. BOCK; ZACH F. JONES; JANE H. BOCK; HOBART M. SMITH
errShare
errSave
Correlation clustering
err2004-07-01
err782
errOAAI
errBansal, N; Blum, A; Chawla, S
errShare
errSave
researcher View more