Summary
Making posters with text descriptions is tricky because it's hard to describe where each part should go using only words. The authors created a new tool called Compo that lets people arrange poster elements directly in space, combining different ways to connect text and images. Compo can work when users build the layout step-by-step or by transforming a high-level request into a detailed plan automatically. They made a way to teach Compo using automatically created training data and tested it against other systems, finding it controls poster composition better while keeping good image quality. This method changes poster making from writing prompts into composing designs directly.
What this means in practice
- •For graphic designers: Create posters by directly arranging visual and textual elements with spatial control instead of describing layouts textually.
- •For marketing teams: Generate high-quality promotional posters from high-level requests that automatically turn into detailed visual layouts.
Authors
Yitong Wang, Fangyun Wei, Jinjing Zhao, Sirui Zhang, Hongyang Zhang, Dong Chen, Bo Dai, Yan Lu
Abstract
Text prompting is an indirect interface for poster generation, requiring users to encode inherently two-dimensional composition intent into a one-dimensional sequence of words. We introduce a Spatial Canvas Interface that enables users to directly compose generation intent in space through four complementary binding types: semantic, identity, text, and pixel, together with Text Specifications for individual elements and global appearance. Based on this interface, we develop Compo, a poster generation model adapted from a pretrained image editing model to understand Spatial Canvas inputs and Text Specifications. Compo supports both direct inference, where users explicitly construct the canvas, and agentic mode, where a high-level request is automatically translated into a planned Spatial Canvas. To train Compo, we develop a scalable pipeline that automatically constructs supervision data for different binding types and their combinations, enabling efficient adaptation without training a specialized poster generator from scratch. We further introduce a benchmark that evaluates adherence to individual binding types and their joint composition. Experiments show that Compo achieves stronger compositional controllability than both general-purpose image generation models and dedicated poster generation systems while maintaining high visual quality. By decoupling intent specification from visual generation, our work shifts poster generation from prompting toward composing.