ZichengDuan/EZIGen

[BMVC 2025] Official implementation for paper EZIGen: Enhancing zero-shot personalized image generation with precise subject encoding and decoupled guidance

39
/ 100
Emerging

This tool helps creative professionals and digital artists generate new images that precisely feature a specific person, animal, or object from a reference photo, combined with a custom text description. It takes a subject image and a text prompt (e.g., 'a dog in police outfit') to produce a brand-new image where the subject maintains their exact identity while appearing in the described scene or style. It's designed for anyone who needs to generate highly personalized images without extensive training data.

107 stars. No commits in the last 6 months.

Use this if you need to create images where a specific subject from a reference photo appears in a completely new, text-described context, or if you want to edit an existing image by replacing or modifying a foreground subject while preserving its identity.

Not ideal if you need to generate images without any reference subject, or if you require extensive control over global image composition rather than focusing on precise subject integration.

digital-art personalized-content image-generation creative-design visual-content-creation
Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 12 / 25

How are scores calculated?

Stars

107

Forks

11

Language

Python

License

Apache-2.0

Last pushed

Aug 21, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/generative-ai/ZichengDuan/EZIGen"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.