000200180 001__ 200180
000200180 005__ 20190316235933.0
000200180 037__ $$aSTUDENT
000200180 245__ $$aLearning Separable Filters with Shared Parts
000200180 269__ $$a2013
000200180 260__ $$c2013
000200180 336__ $$aStudent Projects
000200180 520__ $$aLearned image features can provide great accuracy in many Computer Vision tasks. However, when the convolution filters used to learn image features are numerous and not separable, feature extraction becomes computationally demanding and impractical to use in real-world situations. In this thesis work, a method for learning a small number of separable filters to approximate an arbitrary non-separable filter bank is developed. In this approach, separable filters are learned by grouping the arbitrary filters into a tensor and optimizing a tensor decomposition problem. The separable filter learning with tensor decomposition is general and can be applied to generic filter banks to reduce the computational burden of convolutions without a loss in performance. Moreover, the proposed approach is orders of magnitude faster than the approach of a recent studies based on l1-norm minimization.
000200180 6531_ $$aConvolutional sparse coding
000200180 6531_ $$afilter learning
000200180 6531_ $$afeatures extraction
000200180 6531_ $$aseparable convolution
000200180 6531_ $$asegmentation of linear structures
000200180 6531_ $$aimage denoising
000200180 6531_ $$aconvolutional neural networks
000200180 6531_ $$atensor decomposition
000200180 700__ $$0247609$$g211045$$aTekin, Bugra
000200180 720_2 $$aFua, Pascal$$edir.$$g112366$$0240252
000200180 720_2 $$aLepetit, Vincent$$edir.$$g149007$$0240235
000200180 8564_ $$uhttps://infoscience.epfl.ch/record/200180/files/ms_thesis_tekin_1.pdf$$zn/a$$s2627340$$yn/a
000200180 909C0 $$xU10659$$0252087$$pCVLAB
000200180 909CO $$qGLOBAL_SET$$pIC$$ooai:infoscience.tind.io:200180
000200180 917Z8 $$x211045
000200180 917Z8 $$x211045
000200180 917Z8 $$x211045
000200180 917Z8 $$x211045
000200180 937__ $$aEPFL-STUDENT-200180
000200180 973__ $$aEPFL
000200180 980__ $$bMASTERS$$aSTUDENT