Cross-media visual communication style unified generation system and rendering method

Abstract

A cross-media visual communication style unified generation system and a rendering method, relating to the technical field of visual communication. The system comprises a style feature decoupling module, a media attribute perception module, an anchor style constraint module, a cross-media rendering engine, and a consistency quantitative evaluation module; the system decomposes source visual samples through a dual-branch attention distillation algorithm, acquires physical parameters of a target medium to construct adaptation rules, and relies on anchor constraints to achieve layered rendering and closed-loop verification, thereby solving the problems of cross-media style distortion, poor consistency, and inefficient manual debugging in the prior art. The whole process operates automatically, which not only ensures the unification of core style anchors but also adapts to characteristics of different media, greatly improving the efficiency of cross-media visual generation and the effectiveness of communication, the system is suitable for multi-scenario cross-terminal visual output.