SAMOT: State-Aware Step Modulation and Optimal Transport Matching for Audio-Visual Instance Segmentation
A framework with Adaptive Dynamic Step Modulation (ADSM) and Optimal Transport-based Matching Modulation (OT-MM) is proposed, which adaptively modulates Mamba step sizes using temporal variation, cross-modal discrepancy, and historical context, balancing rapid response to modality-state changes with stable long-range m...