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

Text-driven person image generation is an emerging and challenging task incross-modality image generation. Controllable person image generation promotesa wide range of applications such as digital human interaction and virtualtry-on. However, previous methods mostly employ single-modality information asthe prior condition (e.g. pose-guided person image generation), or utilize thepreset words for text-driven human synthesis. Introducing a sentence composedof free words with an editable semantic pose map to describe person appearanceis a more user-friendly way. In this paper, we propose HumanDiffusion, acoarse-to-fine alignment diffusion framework, for text-driven person imagegeneration. Specifically, two collaborative modules are proposed, the StylizedMemory Retrieval (SMR) module for fine-grained feature distillation in dataprocessing and the Multi-scale Cross-modality Alignment (MCA) module forcoarse-to-fine feature alignment in diffusion. These two modules guarantee thealignment quality of the text and image, from image-level to feature-level,from low-resolution to high-resolution. As a result, HumanDiffusion realizesopen-vocabulary person image generation with desired semantic poses. Extensiveexperiments conducted on DeepFashion demonstrate the superiority of our methodcompared with previous approaches. Moreover, better results could be obtainedfor complicated person images with various details and uncommon poses.

HumanDiffusion: a Coarse-to-Fine Alignment Diffusion Framework for Controllable Text-Driven Person Image Generation