Finally, cross-domain knowledge distillation and pseudolabel filtering enforce consistency regularization based on category consistency priors between unlabeled measured and simulated-like samples, ...