Title: A FAST VISUAL GREEDY SPARSE APPROXIMATION VIA IMPROVED LEARNED DICTIONARY
Abstract:Orthogonal Matching Pursuit (OMP) is an effective solution to sparse approximation based on redundant dictionary, but there are plenty of matrix-vector multiplications in order to seek for the most ma...Orthogonal Matching Pursuit (OMP) is an effective solution to sparse approximation based on redundant dictionary, but there are plenty of matrix-vector multiplications in order to seek for the most matching atom. This full search scheme can consume much CPU time, especially for high dimension data such as image. Although some existing schemes proposed some accelerating methods by exploiting the property of atoms in parametric dictionary, they cannot be extended to other dictionaries, especially learned dictionary. Considering statistical nature of atoms in learned dictionary, this paper proposes a novel method which utilizes modified K-mean algorithm to cluster all the atoms in the learned dictionary while sufficiently considering the inner product operation between every two atoms, determines the class in which the desired atom locates and finally find it in the class. Some analyses and experiments have shown the success of the method. In addition, this paper gives some empirical analysis of the effect of related parameters of indexed dictionary.Read More
Publication Year: 2010
Publication Date: 2010-10-01
Language: en
Type: article
Indexed In: ['crossref']
Access and Citation
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot