Fast image classification method based on multi-level approximate dictionary learning

doi:10.1016/j.dsp.2026.106297

Fast image classification method based on multi-level approximate dictionary learning

2026-05-28

0

PRE

AI

DOI：10.1016/j.dsp.2026.106297

Original

Original request for help

Save

Abstract

En 中文

In this paper, we propose a fast image classification method based on multi-level approximation dictionary learning (MA-DL) to solve the problem of reduced classification performance due to the diversification of image features in practical application scenarios. First, the category label information of the training samples is integrated into the dictionary model to improve the discriminability of the dictionary. Secondly, the Laplacian eigenmap (LE) is used to map the original data to a new low-dimensional feature space to retain the local feature structure of the sample. By minimizing the intra-class deviation between sparse coefficients, the influence of outliers is weakened to suppress intra-class variation. Then, the classification error constraint is introduced to solve the similarity optimal value by minimizing the error between the classification label vector and the linear mapping vector of the sparse coefficients. Through the optimization iteration of multiple parameters, multi-level fast approximation is achieved to improve the convergence speed. Finally, a linear classifier is used for image classification based on the dictionary learning model. Experimental results show that the MA-DL method effectively realizes image classification in different scenarios and effectively improves the classification efficiency.

Keywords:

multi-level approximation

dictionary learning

image classification

sparse coefficients

Laplacian eigenmap

Fast image classification method based on multi-level approximate dictionary learning

Abstract

Journal

Organization