Compositional Text-to-Image Synthesis with Attention Map Control of Diffusion Models
This addresses the issue of poor compositional capabilities in text-to-image synthesis for users needing more accurate and reliable image generation from complex prompts, though it is incremental as it builds on existing diffusion models.
The paper tackles the problem of semantic misalignment in text-to-image diffusion models, such as attribute leakage and missing entities, by proposing an attention mask control strategy based on predicted object boxes, which improves compositional accuracy and can be integrated as a plugin into existing generators.
Recent text-to-image (T2I) diffusion models show outstanding performance in generating high-quality images conditioned on textual prompts. However, they fail to semantically align the generated images with the prompts due to their limited compositional capabilities, leading to attribute leakage, entity leakage, and missing entities. In this paper, we propose a novel attention mask control strategy based on predicted object boxes to address these issues. In particular, we first train a BoxNet to predict a box for each entity that possesses the attribute specified in the prompt. Then, depending on the predicted boxes, a unique mask control is applied to the cross- and self-attention maps. Our approach produces a more semantically accurate synthesis by constraining the attention regions of each token in the prompt to the image. In addition, the proposed method is straightforward and effective and can be readily integrated into existing cross-attention-based T2I generators. We compare our approach to competing methods and demonstrate that it can faithfully convey the semantics of the original text to the generated content and achieve high availability as a ready-to-use plugin. Please refer to https://github.com/OPPOMente-Lab/attention-mask-control.