Text this: Interpretable multimodal emotion recognition using hybrid fusion of speech and image data.