SOW uses MLLMs and attention to selectively control unidirectional diffusion for pixel-level fidelity and contextual coherence in text-vision-to-image tasks.
Attend-and- excite: Attention-based semantic guidance for text-to-image diffusion models,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
SOWing Information: Cultivating Contextual Coherence with MLLMs in Image Generation
SOW uses MLLMs and attention to selectively control unidirectional diffusion for pixel-level fidelity and contextual coherence in text-vision-to-image tasks.