Object-Attribute Binding in Text-to-Image Generation: Evaluation and Control
This addresses a key limitation in text-to-image generation for users needing accurate visual representations from text, though it is incremental as it builds on existing diffusion models.
The paper tackles the problem of attribute-object binding in text-to-image generation, where current diffusion models often misassign attributes to objects, and proposes a method using focused cross-attention and DisCLIP embeddings to improve this, showing substantial gains on multiple datasets.
Current diffusion models create photorealistic images given a text prompt as input but struggle to correctly bind attributes mentioned in the text to the right objects in the image. This is evidenced by our novel image-graph alignment model called EPViT (Edge Prediction Vision Transformer) for the evaluation of image-text alignment. To alleviate the above problem, we propose focused cross-attention (FCA) that controls the visual attention maps by syntactic constraints found in the input sentence. Additionally, the syntax structure of the prompt helps to disentangle the multimodal CLIP embeddings that are commonly used in T2I generation. The resulting DisCLIP embeddings and FCA are easily integrated in state-of-the-art diffusion models without additional training of these models. We show substantial improvements in T2I generation and especially its attribute-object binding on several datasets.\footnote{Code and data will be made available upon acceptance.