Using the MGB-2 challenge data for creating a new multimodal Dataset for speaker role recognition in Arabic TV Broadcasts.

Mohamed Lazhar Bellagha, Mounir Zrigui

Browse the full KES paper archive.