Exploiting modality-invariant feature for robust multimodal emotion recognition with missing modalities

Open in new window