An orbicularis oris, buccinator, zygomaticus, and risorius muscle contraction classification for lip-reading during speech using sEMG signals on multi-channels.