Face-StyleSpeech: Enhancing Zero-shot Speech Synthesis from Face Images with Improved Face-to-Speech Mapping

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Kang, Minki, Han, Wooseok, Yang, Eunho
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916543920603136
author Kang, Minki
Han, Wooseok
Yang, Eunho
author_facet Kang, Minki
Han, Wooseok
Yang, Eunho
contents Generating speech from a face image is crucial for developing virtual humans capable of interacting using their unique voices, without relying on pre-recorded human speech. In this paper, we propose Face-StyleSpeech, a zero-shot Text-To-Speech (TTS) synthesis model that generates natural speech conditioned on a face image rather than reference speech. We hypothesize that learning entire prosodic features from a face image poses a significant challenge. To address this, our TTS model incorporates both face and prosody encoders. The prosody encoder is specifically designed to model speech style characteristics that are not fully captured by the face image, allowing the face encoder to focus on extracting speaker-specific features such as timbre. Experimental results demonstrate that Face-StyleSpeech effectively generates more natural speech from a face image than baselines, even for unseen faces. Samples are available on our demo page.
format Preprint
id arxiv_https___arxiv_org_abs_2311_05844
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Face-StyleSpeech: Enhancing Zero-shot Speech Synthesis from Face Images with Improved Face-to-Speech Mapping
Kang, Minki
Han, Wooseok
Yang, Eunho
Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
Multimedia
Sound
Audio and Speech Processing
Generating speech from a face image is crucial for developing virtual humans capable of interacting using their unique voices, without relying on pre-recorded human speech. In this paper, we propose Face-StyleSpeech, a zero-shot Text-To-Speech (TTS) synthesis model that generates natural speech conditioned on a face image rather than reference speech. We hypothesize that learning entire prosodic features from a face image poses a significant challenge. To address this, our TTS model incorporates both face and prosody encoders. The prosody encoder is specifically designed to model speech style characteristics that are not fully captured by the face image, allowing the face encoder to focus on extracting speaker-specific features such as timbre. Experimental results demonstrate that Face-StyleSpeech effectively generates more natural speech from a face image than baselines, even for unseen faces. Samples are available on our demo page.
title Face-StyleSpeech: Enhancing Zero-shot Speech Synthesis from Face Images with Improved Face-to-Speech Mapping
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
Multimedia
Sound
Audio and Speech Processing
url https://arxiv.org/abs/2311.05844