RegBN: Batch Normalization of Multimodal Data with Regularization

However, the integration of heterogeneous multimodal data poses a significant challenge, as confounding effects and dependencies among such heterogeneous data sources introduce unwanted variability and bias, leading to suboptimal performance of multimodal models. Therefore, it becomes crucial to normalize the low-or high-level features extracted from data modalities before their fusion takes place.