Searcharxiv⌕ Search

arXiv subjects

Hailong Ye

Publications and source records attributed to Hailong Ye.

3 recordsLinked to original sources

Test-Time Selective Adaptation for Cross-Modal UAV Object Tracking with Token Shift

Cross-modal unmanned aerial vehicle (UAV) object tracking faces modality-induced appearance discrepancies and abrupt target-position shifts that can arise from modality switching, UAV motion, and inter-sensor offsets. Although existing methods model modality and spatial variations, they primarily rely on offline representation learning and offer limited localization-oriented adaptation during testing. To address these challenges, we propose Test-Time Selective Adaptation with Token Shift (TSAT), which derives spatial and temporal adaptation signals from unlabeled test sequences for online optimization. Specifically, we design a controllable token shift auxiliary task, which imposes explicit spatial correspondence between predefined token translations and resultant localization offsets, generating stable annotation-free spatial adaptation signals. Furthermore, we introduce a temporal backward prediction auxiliary task to exploit frame-wise temporal dependencies, effectively mitigating cumulative localization errors in continuous tracking. To suppress interference from unreliable tracking samples, we develop a reliability-aware update gating strategy to screen valid adaptation signals and stabilize online parameter updates. Experiments on CM-UOT demonstrate that TSAT achieves state-of-the-art performance against existing test-time adaptation methods, exhibiting superior robustness under frequent modality switching and severe spatial displacement. Experiments across tracking frameworks and datasets further validate the effectiveness and generalization capability of TSAT.

cs.CV↗

Global Dynamics of Nonlocal Diffusion Systems on Time-Varying Domains

We propose a class of nonlocal diffusion systems on time-varying domains, and fully characterize their asymptotic dynamics in the asymptotically fixed, time-periodic and unbounded cases. The kernel is not necessarily symmetric or compactly supported, provoking anisotropic diffusion or convective effects. Due to the nonlocal diffusion on time-varying domains in our systems, some significant challenges arise, such as the lack of regularizing effects of the semigroup generated by the nonlocal operator, as well as the time-dependent inherent coupling structure in kernel. By investigating a general nonautonomous nonlocal diffusion system in the space of bounded and measurable functions, we establish a comprehensive and unified framework to rigorously examine the threshold dynamics of the original system on asymptotically fixed and time-periodic domains. In the case of an asymptotically unbounded domain, we introduce a key auxiliary function to separate vanishing coefficients from nonlocal diffusions. This enables us to construct appropriate sub-solutions and derive the global threshold dynamics via the comparison principle. The findings may be of independent interest and the developed techniques, which do not rely on the existence of the principal eigenvalue, are expected to find further applications in the related nonlocal diffusion problems. We also conduct numerical simulations based on a practical model to illustrate our analytical results.

math.AP↗

Bounded weak and strong time periodic solutions to a three-dimensional chemotaxis-Stokes model with porous medium diffusion

In this paper, we study the time periodic problem to a three-dimensional chemotaxis-Stokes model with porous medium diffusion $Δn^m$ and inhomogeneous mixed boundary conditions. By using a double-level approximation method and some iterative techniques, we obtain the existence and time-space uniform boundedness of weak time periodic solutions for any $m>1$. Moreover, we improve the regularity for $m\le\frac{4}{3}$ and show that the obtained periodic solutions are in fact strong periodic solutions.

math.AP↗