TextDiffuser-2: Unleashing the Power of Language Models for Text Rendering

TL;DR

TextDiffuser-2 leverages large language models within diffusion frameworks to improve text rendering by enabling automated layout planning, diverse style generation, and flexible modifications, validated through extensive experiments and user studies.

Contribution

The paper introduces a novel method that integrates large language models with diffusion models for automated, flexible, and diverse text rendering.

Findings

Achieves more rational text layouts and diverse styles

Outperforms previous methods in flexibility and automation

Validated by human and GPT-4V user studies

Abstract

The diffusion model has been proven a powerful generative model in recent years, yet remains a challenge in generating visual text. Several methods alleviated this issue by incorporating explicit text position and content as guidance on where and what text to render. However, these methods still suffer from several drawbacks, such as limited flexibility and automation, constrained capability of layout prediction, and restricted style diversity. In this paper, we present TextDiffuser-2, aiming to unleash the power of language models for text rendering. Firstly, we fine-tune a large language model for layout planning. The large language model is capable of automatically generating keywords for text rendering and also supports layout modification through chatting. Secondly, we utilize the language model within the diffusion model to encode the position and texts at the line level. Unlike previous methods that employed tight character-level guidance, this approach generates more diverse text images. We conduct extensive experiments and incorporate user studies involving human participants as well as GPT-4V, validating TextDiffuser-2's capacity to achieve a more rational text layout and generation with enhanced diversity. The code and model will be available at \url{https://aka.ms/textdiffuser-2}.