English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The generation of emotional talking faces from a single portrait imageremains a significant challenge. The simultaneous achievement of expressiveemotional talking and accurate lip-sync is particularly difficult, asexpressiveness is often compromised for the accuracy of lip-sync. As widelyadopted by many prior works, the LSTM network often fails to capture thesubtleties and variations of emotional expressions. To address thesechallenges, we introduce DREAM-Talk, a two-stage diffusion-based audio-drivenframework, tailored for generating diverse expressions and accurate lip-syncconcurrently. In the first stage, we propose EmoDiff, a novel diffusion modulethat generates diverse highly dynamic emotional expressions and head poses inaccordance with the audio and the referenced emotion style. Given the strongcorrelation between lip motion and audio, we then refine the dynamics withenhanced lip-sync accuracy using audio features and emotion style. To this end,we deploy a video-to-video rendering module to transfer the expressions and lipmotions from our proxy 3D avatar to an arbitrary portrait. Both quantitativelyand qualitatively, DREAM-Talk outperforms state-of-the-art methods in terms ofexpressiveness, lip-sync accuracy and perceptual quality.

DREAM-Talk: Diffusion-based Realistic Emotional Audio-driven Method for Single Image Talking Face Generation