(Translated by https://www.hiragana.jp/)
[2406.08850] COVE: Unleashing the Diffusion Feature Correspondence for Consistent Video Editing