Structured sparse representation and low-dimension embedding combined dictionary learning method

Abstract

The invention discloses a structured sparse representation and low-dimension embedding combined dictionary learning method. Dictionary construction and dimension-reduced matrix projection are carriedout in parallel alternatively, a sparse representation coefficient matrix forced to include a block diagonal structure in a low-dimension projection space is used to enhance inter-type incoherence ofa dictionary, intra-type coherence of the dictionary is maintained via low-rank performance of a representative coefficient in a type sub-dictionary, dictionary construction and projection learning prompt each other to main the sparse structure of data fully, and a coding coefficient with a higher type discriminating capability is coded. The problems that the coded representation coefficient is low in discriminating and distinguishing capabilities in an existing dictionary learning method in the prior art due to high maintaining characteristic of training sample data and lack of a dictionary with strictly constraint of the block diagonal structure.

Description

technical field [0001] The invention belongs to the technical field of digital image processing, and in particular relates to a dictionary learning method combining structured sparse representation and low-dimensional embedding. Background technique [0002] The core idea of ​​sparse representation is mainly based on the objective fact that many signals in nature can be represented or coded by a linear combination of only a few dictionary entries in an oversaturated dictionary. The most critical issue in sparse representation research is the construction of dictionaries with strong representation capabilities. At present, sparse representation technology is widely used in many application fields, such as image classification, face recognition and human action recognition. [0003] Dictionary learning is dedicated to learning an optimal dictionary from training samples in order to better represent or encode a given signal or feature. For classification recognition based on ...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a dictionary learning method that combines structured sparse representation and low-dimensional embedding, which solves the problem of the high-dimensional characteristics of training sample data and the lack of strict block diagonalization structure in the dictionary learning method existing in the prior art. Constrained dictionary makes the coded representation coefficient class discrimination ability weak and the discrimination is not strong

Images