DreamPoster: A Unified Framework for Image-Conditioned Generative Poster Design

TL;DR

DreamPoster is a comprehensive framework that generates high-quality, content-faithful posters from images and text prompts, supporting flexible layouts and resolutions with superior performance over existing methods.

Contribution

It introduces a unified Text-to-Image poster generation model, a systematic dataset annotation pipeline, and a progressive training strategy for multi-task poster synthesis.

Findings

Achieves an 88.55% usability rate, outperforming GPT-4o and SeedEdit3.0.

Supports flexible resolution and layout outputs.

Demonstrates high content fidelity and generation quality.

Abstract

We present DreamPoster, a Text-to-Image generation framework that intelligently synthesizes high-quality posters from user-provided images and text prompts while maintaining content fidelity and supporting flexible resolution and layout outputs. Specifically, DreamPoster is built upon our T2I model, Seedream3.0 to uniformly process different poster generating types. For dataset construction, we propose a systematic data annotation pipeline that precisely annotates textual content and typographic hierarchy information within poster images, while employing comprehensive methodologies to construct paired datasets comprising source materials (e.g., raw graphics/text) and their corresponding final poster outputs. Additionally, we implement a progressive training strategy that enables the model to hierarchically acquire multi-task generation capabilities while maintaining high-quality generation. Evaluations on our testing benchmarks demonstrate DreamPoster's superiority over existing methods, achieving a high usability rate of 88.55\%, compared to GPT-4o (47.56\%) and SeedEdit3.0 (25.96\%). DreamPoster will be online in Jimeng and other Bytedance Apps.