AI RESEARCH
Beyond ZOH: Advanced Discretization Strategies for Vision Mamba
arXiv CS.CV
•
ArXi:2604.20606v1 Announce Type: new Vision Mamba, as a state space model (SSM), employs a zero-order hold (ZOH) discretization, which assumes that input signals remain constant between sampling instants. This assumption degrades temporal fidelity in dynamic visual environments and constrains the attainable accuracy of modern SSM-based vision models.