Refining Text-to-Image Generation: Towards Accurate Training-Free Glyph-Enhanced Image Generation
This work addresses a specific issue in text-to-image generation for applications requiring accurate visual text, though it is incremental as it builds on existing two-stage approaches.
The paper tackles the problem of inaccurate text spelling in text-to-image generation by introducing a training-free framework to enhance glyph-controlled models, demonstrating improvements such as over 23% and 13.5% in OCR word F1 scores on new benchmarks for lengthy and complex text.
Over the past few years, Text-to-Image (T2I) generation approaches based on diffusion models have gained significant attention. However, vanilla diffusion models often suffer from spelling inaccuracies in the text displayed within the generated images. The capability to generate visual text is crucial, offering both academic interest and a wide range of practical applications. To produce accurate visual text images, state-of-the-art techniques adopt a glyph-controlled image generation approach, consisting of a text layout generator followed by an image generator that is conditioned on the generated text layout. Nevertheless, our study reveals that these models still face three primary challenges, prompting us to develop a testbed to facilitate future research. We introduce a benchmark, LenCom-Eval, specifically designed for testing models' capability in generating images with Lengthy and Complex visual text. Subsequently, we introduce a training-free framework to enhance the two-stage generation approaches. We examine the effectiveness of our approach on both LenCom-Eval and MARIO-Eval benchmarks and demonstrate notable improvements across a range of evaluation metrics, including CLIPScore, OCR precision, recall, F1 score, accuracy, and edit distance scores. For instance, our proposed framework improves the backbone model, TextDiffuser, by more than 23\% and 13.5\% in terms of OCR word F1 on LenCom-Eval and MARIO-Eval, respectively. Our work makes a unique contribution to the field by focusing on generating images with long and rare text sequences, a niche previously unexplored by existing literature