Multi-layer graph constraint dictionary pair learning for image classification

doi:10.1016/j.jvcir.2025.104638

Multi-layer graph constraint dictionary pair learning for image classification

2025-11-06

0

PRE

AI

DOI：10.1016/j.jvcir.2025.104638

Original

Original request for help

Save

Abstract

En 中文

• A powerful architecture, called the multi-layer graph constraint dictionary pair learning (MGDPL) with structure graph constraint, is proposed for image classification. • The MGDPL seamlessly integrates multi-layer dictionary pair learning, structure graph constraint, and discriminative sparse representations into a unified model. • The MGDPL uses a multi-layer structure dictionary learning framework to deliver hierarchical learning of dictionary pairs. • The MGDPL imposes the structure graph constraint on the sub-sparse representations to ensure the discriminative ability of near-neighbor graphs. • The discriminative multi-layer graph regularized constraint term can ensure high intra-class compactness and inter-class separation in the data reconstruction space.

Multi-layer graph constraint dictionary pair learning for image classification

Abstract

Journal

Organization