An iterative text-image plan generation framework with pseudo-PDDL visual feedback improves multimodal consistency and visual coherence across Mistral-7B, Gemini-1.5, and GPT-4o.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Enhance Multimodal Consistency and Coherence for Text-Image Plan Generation
An iterative text-image plan generation framework with pseudo-PDDL visual feedback improves multimodal consistency and visual coherence across Mistral-7B, Gemini-1.5, and GPT-4o.