domain generalization aims to solve the challenge of out-of-distribution (OOD) generalization by leveraging common knowledge learned from multiple training domains to generalize to unseen test domains. To accurat