IEEE Transactions on Signal Processing · 2007 · 19 citations · 19 references
ComputingComputer ScienceMathematicsAlgorithmsInformation TheoryMachine LearningApproximation AlgorithmsData CompressionDigital Signal ProcessingDimensionality ReductionPattern RecognitionCompressed SensingDictionary LearningGreedy ApproximationGreedy HeuristicsSparse CodingSparse ReconstructionSparse Signal ProcessingSparse Signal Representation
Recently, there has been an intense activity in the field of sparse approximations with redundant dictionaries, largely motivated by the practical performances of algorithms such as matching pursuit (MP) and basis pursuit (BP). However, most of the theoretical results obtained so far are valid only for the restricted class of incoherent dictionaries. This paper investigates a new class of overcomplete dictionaries, called block incoherent dictionaries, where coherence can be arbitrarily big. We show that a simple greedy algorithm can correctly identify stable subdictionaries (called blocks) and demonstrate how one can use the extra coherence freedom for approximation purposes.
19
Atomic Decomposition by Basis Pursuit
Scott Shaobing Chen, David L. Donoho, Michael A. Saunders · SIAM Review · 2001 · 5.1K citations