Multi-Microphone and Multi-Modal Emotion Recognition in Reverberant Environments