Virtual Immortality: Reanimating Characters from TV Shows

摘要

The objective of this work is to build virtual talking avatars of characters fully automatically from TV shows. Prom this unconstrained data, we show how to capture a character's style of speech, visual appearance and language in an effort to construct an interactive avatar of the person and effectively immortalize them in a computational model. We make three contributions (ⅰ) a complete framework for producing a generative model of the audiovisual and language of characters from TV shows; (ⅱ) a novel method for aligning transcripts to video using the audio; and (ⅲ) a fast audio segmentation system for silencing non-spoken audio from TV shows. Our framework is demonstrated using all 236 episodes from the TV series Friends (≈97 h of video) and shown to generate novel sentences as well as character specific speech and video.