HumanAIGC/EMO

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

7594 stars · 928 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

EMO (Emote Portrait Alive) is a research codebase from Alibaba's Institute for Intelligent Computing that generates expressive talking portrait videos from a single reference image and an audio clip, using an audio2video diffusion model under weak conditions. It was published at ECCV 2024 and is primarily a paper companion release.

Use cases

• generate a talking head video from a photo and audio
• animate a portrait image to match speech audio
• create expressive avatar videos with audio-driven diffusion
• reproduce ECCV 2024 audio2video portrait research
• build lip-synced character videos from voice recordings

When to choose

• you need state-of-the-art audio-driven portrait video generation for research
• you want to study or extend the EMO diffusion approach
• you have GPU resources and are comfortable with research-grade code

When to avoid

• you need a production-ready tool with a stable API or license
• you lack a GPU or cannot run heavy diffusion inference
• you need commercial usage rights - the repo has no license
• you want a polished end-user application rather than research code

Facets

library · maturity experimental

machine-learning deep-learning video-processing audio-processing computer-vision speech-recognition artificial-intelligence deep-learning computer-vision python audio2video diffusion-model talking-head portrait-animation research-code eccv-2024 no-license video audio gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="HumanAIGC/EMO")