Relational graph-driven differential denoising and diffusion attention fusion for multimodal conversation emotion recognition
arXiv:2603.25752v1 Announce Type: new
Abstract: In real-world scenarios, audio and video signals are often subject to environmental noise and limited acquisition conditions, resulting in extracted features containing excessive noise. Furthermore, ther…