Abstract: A multimodal system, device, and method for capturing, processing, and overlaying subcutaneous vasculature and anatomical data onto extended reality (XR) displays, enabling hands-free visualization for clinical procedures. The system integrates one or more imaging modalities (e.g., ultrasound, near-infrared, infrared, photoacoustic imaging, optical coherence tomography, transillumination, hyperspectral imaging, laser speckle contrast) with AI-assisted processing pipelines to create real-time or pre-scanned overlays. The processed data is visualized via XR devices, including augmented reality (AR), mixed reality (MR), projection-based AR, and virtual reality (VR) systems. The system supports real-time segmentation, depth estimation, predictive analytics, adaptive overlays, and patient-specific recommendations, providing improved accuracy for vascular access, oncology, cardiology, neonatal care, trauma management, telemedicine, and other clinical workflows.

Abstract: A computer-implemented system and method for predictive scene-based user interface orchestration are disclosed. A plurality of context parameters, including behavioral, environmental, temporal, task-related, social, emotional, and device-ecosystem data, are acquired from sensors and applications. A context vector is generated comprising parameter values with confidence scores and temporal-decay values, and a context score is computed by applying a weighted aggregation. One or more context states are predicted using a machine learning model, and candidate interface scenes comprising layout structures, interaction flows, and component hierarchies are pre-generated. A transition policy is applied to select a candidate scene, which is rendered with continuity preservation and cinematic transitions that maintain user mental models and application state. Upon detection of unsatisfactory performance, rollback to a prior scene is executed.