Multimodal diffusion framework for collaborative text image audio generation and applications

Abstract This paper presents a novel framework for collaborative generation across text, image, and audio modalities using an enhanced diffusion model architecture. We introduce a Hierarchical Cross-modal Alignment Network that establishes unified representations while preserving modality-specific c...

Full description

Saved in:
Bibliographic Details
Main Authors: Junhua Wang, Ouya Zhang, Yuan Jiang
Format: Article
Language:English
Published: Nature Portfolio 2025-07-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-05794-4
Tags: Add Tag
No Tags, Be the first to tag this record!