Two Heads Are Better Than One: Audio-Visual Speech Error Correction with Dual Hypotheses

Re-assign community