Domain generalization image classification method based on low-rank constraint local regression

Abstract

The invention relates to a domain generalization image classification method based on low-rank constraint local regression, and the method comprises the steps: (1), obtaining a source domain, and decomposing the source domain into a plurality of sub-domains; (2) training according to each sub-domain to obtain a classifier; step (3), constructing a low-rank constraint local regression framework according to the classifier obtained by training each sub-domain; and step (4), classifying the image based on the low-rank constraint local regression framework. The method can effectively classify the images, and can be applied to domain self-adaption.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a domain generalization image classification method based on low-rank constrained local regression. Background technique [0002] Humans are good at using past experience to solve problems quickly in new environments. Likewise, we expect machines to also build robust models under different conditions (domains) to solve many new tasks. For example, we want computer vision systems to be able to recognize objects immediately when the camera environment changes, without retraining. Methods to solve this problem can be classified into domain adaptation (DA) and domain generalization (DG). Domain Adaptation (DA) methods have been successfully applied to a wide range of vision applications by reducing the distribution gap in different but related domains and using source and target domain samples to train a target domain-specific model. As a related research problem, domain g...

AI Technical Summary

Problems solved by technology

However, how to efficiently divide training samples into multiple clusters to discover feature latent domains is a challenging task

Images