Search papers, labs, and topics across Lattice.
This paper introduces MambaTrack, a new RGB-Event object tracking framework that uses a Dynamic State Space Model to address limitations in existing Vision Mamba-based methods. The core innovation is an event-adaptive state transition mechanism that modulates the state transition matrix based on event stream density, allowing for differentiated modeling of sparse and dense event flows. Additionally, a Gated Projection Fusion module is introduced to project RGB features into the event feature space, using adaptive gates derived from event density and RGB confidence scores to control fusion intensity. MambaTrack achieves state-of-the-art performance on FE108 and FELT datasets while maintaining a lightweight design suitable for real-time embedded deployment.
By dynamically adjusting state transitions based on event sparsity, MambaTrack achieves state-of-the-art RGB-Event object tracking, outperforming methods stuck with static transition matrices.
Existing Vision Mamba-based RGB-Event(RGBE) tracking methods suffer from using static state transition matrices, which fail to adapt to variations in event sparsity. This rigidity leads to imbalanced modeling-underfitting sparse event streams and overfitting dense ones-thus degrading cross-modal fusion robustness. To address these limitations, we propose MambaTrack, a multimodal and efficient tracking framework built upon a Dynamic State Space Model(DSSM). Our contributions are twofold. First, we introduce an event-adaptive state transition mechanism that dynamically modulates the state transition matrix based on event stream density. A learnable scalar governs the state evolution rate, enabling differentiated modeling of sparse and dense event flows. Second, we develop a Gated Projection Fusion(GPF) module for robust cross-modal integration. This module projects RGB features into the event feature space and generates adaptive gates from event density and RGB confidence scores. These gates precisely control the fusion intensity, suppressing noise while preserving complementary information. Experiments show that MambaTrack achieves state-of-the-art performance on the FE108 and FELT datasets. Its lightweight design suggests potential for real-time embedded deployment.
