000196542 001__ 196542
000196542 005__ 20190316235839.0
000196542 022__ $$a0920-5691
000196542 0247_ $$2doi$$a10.1007/s11263-014-0784-7
000196542 037__ $$aARTICLE
000196542 245__ $$aDictionary learning for fast classification based on soft-thresholding
000196542 269__ $$a2015
000196542 260__ $$bSpringer$$c2015$$aDordrecht
000196542 300__ $$a16
000196542 336__ $$aJournal Articles
000196542 520__ $$aClassifiers based on sparse representations have recently been shown to provide excellent results in many visual recognition and classification tasks. However, the high cost of computing sparse representations at test time is a major obstacle that limits the applicability of these methods in large-scale problems, or in scenarios where computational power is restricted. We consider in this paper a simple yet efficient alternative to sparse coding for feature extraction. We study a classification scheme that applies the soft-thresholding nonlinear mapping in a dictionary, followed by a linear classifier. A novel supervised dictionary learning algorithm tailored for this low complexity classification architecture is proposed. The dictionary learning problem, which jointly learns the dictionary and linear classifier, is cast as a difference of convex (DC) program and solved efficiently with an iterative DC solver. We conduct experiments on several datasets, and show that our learning algorithm that leverages the structure of the classification problem outperforms generic learning procedures. Our simple classifier based on soft-thresholding also competes with the recent sparse coding classifiers, when the dictionary is learned appropriately. The adopted classification scheme further requires less computational time at the testing stage, compared to other classifiers. The proposed scheme shows the potential of the adequately trained soft-thresholding mapping for classification and paves the way towards the development of very efficient classification methods for vision problems.
000196542 6531_ $$aDictionary learning
000196542 6531_ $$aSoft-thresholding
000196542 6531_ $$aSparse coding
000196542 6531_ $$aRectifier linear units
000196542 6531_ $$aNeural networks
000196542 700__ $$0246320$$g203034$$uEcole Polytech Fed Lausanne, Signal Proc Lab LTS4, CH-1015 Lausanne, Switzerland$$aFawzi, Alhussein
000196542 700__ $$aDavies, Mike
000196542 700__ $$0241061$$g101475$$uEcole Polytech Fed Lausanne, Signal Proc Lab LTS4, CH-1015 Lausanne, Switzerland$$aFrossard, Pascal
000196542 773__ $$j114$$tInternational Journal Of Computer Vision$$k2-3$$q306-321
000196542 8564_ $$uhttps://infoscience.epfl.ch/record/196542/files/softthresh_classification_final.pdf$$zPreprint$$s788929$$yPreprint
000196542 909C0 $$xU10851$$0252393$$pLTS4
000196542 909CO $$ooai:infoscience.tind.io:196542$$qGLOBAL_SET$$pSTI$$particle
000196542 917Z8 $$x203034
000196542 917Z8 $$x101475
000196542 917Z8 $$x203034
000196542 917Z8 $$x203034
000196542 917Z8 $$x101475
000196542 917Z8 $$x203034
000196542 937__ $$aEPFL-ARTICLE-196542
000196542 973__ $$rREVIEWED$$sPUBLISHED$$aEPFL
000196542 980__ $$aARTICLE