Highlights What are the main findings? Training-free AWIE enhances infrared images via frequency-domain adaptive modulation. Decoupled CGAF prevents feature confusion via independent cross-modal gating. What are the implications of the main findings? AWIE-CGAF achieves 89.5% mAP@0.5 and 51.7 FPS on Jetson AGX Orin with only 13.5 M parameters. Results support real-time IR–D drogue detection on edge platforms. Abstract High-precision drogue localization during terminal guidance is critical to close-range autonomous unmanned aerial vehicle (UAV) docking and hinges on infrared–depth (IR–D) multimodal detection. Yet, deploying such detection on airborne edge computing platforms faces severe challenges due to modal heterogeneity, feature redundancy, and real-time constraints. A lightweight IR–D fusion detection network, termed AWIE-CGAF, is proposed for airborne edge deployment, which integrates frequency-domain, physics-prior-driven input enhancement with decoupled gated attention-based adaptive feature fusion to achieve efficient multimodal detection. A training-free Adaptive Wavelet Image Enhancement (AWIE) module is designed to differentially modulate image structures and details in the frequency domain, improving the signal-to-noise ratio and feature discriminability. Concurrently, a Cross-Gated Attention Fusion (CGAF) module employs decoupled cross-modal attention with independent gating, preserving modality-specific features while dynamically selecting complementary information, mitigating redundancy and feature contamination. Experiments on the self-constructed Drogue Infrared–Depth (DIRD) dataset showed that AWIE-CGAF achieved 89.5% mAP@0.5 and 58.2% mAP@0.5:0.95 with 13.5 M parameters, while maintaining real-time inference at 51.7 FPS on a Jetson AGX Orin edge platform. Among the evaluated methods, the proposed framework achieved the highest detection accuracy while retaining real-time edge inference capability. These results support the feasibility of AWIE-CGAF for resource-constrained IR–D drogue perception.
Bao-Shan Li, Haibo Wang, Dong Cao et al.· Italian National Conference...· 0 citations
This study aimed to investigate the association between baseline cognitive performance and the clinical response to antipsychotics in first-episode schizophrenia (FES) patients.
This study examined 769 patients from a multi-center research cohort. Neurocognition was measured using the MATRICS Consensus Cognitive Battery (MCCB) in schizophrenia at baseline. Clinical response was defined as ≥ 50% reduction in PANSS total score from baseline to week 8. Pearson correlation analysis was performed for preliminary exploratory assessment. Multivariable logistic regression models (response: yes/no) were fitted with each cognitive domain as dependent variable, adjusted confounding factors. To test effect modification, we added interaction terms: (i) cognition × antipsychotic type and (ii) cognition × baseline illness severity. For domains with significant interactions (
P
< 0.05), stratified analyses were performed within each drug group and severity stratum.
Logistic regression analysis demonstrated that all baseline neurocognitive domains were significantly associated with clinical response (All
P
< 0.05). Significant interactions between neurocognitive performance and antipsychotics were observed across all domains except Working Memory and Reasoning/problem solving. Similarly, significant interactions between neurocognition and baseline illness severity were found for all domains except Visual and Verbal learning. Further stratified analyses revealed that baseline cognitive performance was associated to clinical response to amisulpride and risperidone. Additionally, significant associations between neurocognition and clinical response were consistently observed in patients with moderate-to-severe illness severity across all domains, with the exception of Visual and Verbal Learning.
The association between baseline cognitive performance and treatment response is contingent on medication regimen and baseline illness severity.
The study was registered on ClinicalTrials.gov (NCT03451734) with a registration date of January 23, 2018.
Yuanyuan Zhu, Ye Yang, R. Wu et al.· BMC Psychiatry· 0 citations