Aims To assess trends, co-occurrence patterns, and future projections of periodontal disease and edentulism burden among women aged ≥60 years in China and Association of Southeast Asian Nations (ASEAN) countries from 1990 to 2033. Methods Using Global Burden of Disease 2023 data,…
BACKGROUND: Ischemic stroke (IS) is a major cause of mortality and disability globally, with challenges in early diagnosis and prognosis prediction. Dysregulated lipid metabolism is key to IS pathophysiology, but comprehensive profiling of lipid changes during disease progression…
Latent world models improve sample efficiency in continuous control by optimizing policies over imagined latent trajectories, but common neural transitions offer limited direct control over modal persistence and error accumulation in long rollouts. We propose Koopman Dreamer, a D…
Visible-Infrared Person Re-Identification (VI-ReID) operates under a closed-world assumption, where queries and galleries are from heterogeneous modalities. However, in open-world scenarios, both sets are likely to contain homogeneous and heterogeneous modality images. A query ma…
Person re-identification (ReID) serves as a critical component in intelligent surveillance systems, aiming to match identities across disjoint camera networks. While traditional methods primarily rely on single-modal RGB imagery, they are often constrained by environmental challe…
Deep Metric Learning (DML) often struggles with zero-shot generalization because standard objectives inherently capture what co-occurs rather than what causes similarity. Consequently, DML models are vulnerable to shortcut learning driven by two structurally distinct confounders:…
Text-to-image person re-identification (TIPR) retrieves target persons using natural language descriptions. However, existing methods largely overlook resolution variance in real-world surveillance. They characterize cross-resolution TIPR through two coupled failure modes: Eviden…