The Internet of Health Monitoring (IoHM) plays a vital role in emergency medical package delivery (EMPD) by enabling real-time monitoring and transmission of critical health data through interconnected devices in 6G networks. Autonomous aerial vehicles (UAVs) act as aerial base stations (ABSs), facilitating data collection and transmission between GAI-IoHM devices and edge servers. This is crucial for efficient communication in 6G-enabled space-air–ground integrated networks (SAGINs). However, UAVs supporting EMPD face challenges related to limited energy and computational capacity, especially during task offloading to edge servers. To address these constraints, this article proposes the integration of generative artificial intelligence (GAI) into UAVs for intelligent policy learning, enabling adaptive decision-making, real-time diagnostics, and efficient path planning under uncertainty. We present a novel multiagent deep reinforcement learning (MADRL)-based joint optimization framework for cooperative task allocation, trajectory planning, and power management (CTATP) in a 6G-enabled SAGIN architecture. The problem is modeled as a partially observable markov decision process (POMDP) to capture dynamic and uncertain operational conditions. To solve this, we introduce the GAI-based deep deterministic double policy gradient (GAI-DD3PG) algorithm, which leverages a generative actor-network to learn adaptive, energy-efficient control policies from latent action spaces. Simulations in urban emergency scenarios with 10–20 UAVs demonstrate GAI-DD3PG’s efficacy. Compared to proximal policy optimization (PPO), deep deterministic policy gradient (DDPG), multiagent federated reinforcement learning (MAFRL), and Greedy heuristics, GAI-DD3PG achieves a 20% energy reduction, 15% higher delivery success rate, 25% shorter trajectories, and 30% improved resource utilization. These results highlight its potential for reliable EMPD in complex 6G SAGIN environments.

The universality of deep neural networks across different modalities and their generalization capabilities to unseen domains play an essential role in medical image segmentation. The recent segment anything model (SAM) has demonstrated strong adaptability across diverse natural scenarios. However, the huge computational costs, demand for manual annotations as prompts and conflict-prone decoding process of SAM degrade its generalization capabilities in medical scenarios. To address these limitations, we propose a modality-decoupled lightweight SAM for domain-generalized medical image segmentation, named De-LightSAM. Specifically, we first devise a lightweight domain-controllable image encoder (DC-Encoder) that produces discriminative visual features for diverse modalities. Further, we introduce the self-patch prompt generator (SP-Generator) to automatically generate high-quality dense prompt embeddings for guiding segmentation decoding. Finally, we design the query-decoupled modality decoder (QM-Decoder) that leverages a one-to-one strategy to provide an independent decoding channel for every modality, preventing mutual knowledge interference of different modalities. Moreover, we design a multi-modal decoupled knowledge distillation (MDKD) strategy to leverage robust common knowledge to complement domain-specific medical feature representations. Extensive experiments indicate that De-LightSAM outperforms state-of-the-arts in diverse medical imaging segmentation tasks, displaying superior modality universality and generalization capabilities. Especially, De-LightSAM uses only 2.0% parameters compared to SAM-H. The source code is available at https://github.com/xq141839/De-LightSAM.

The Segment Anything Model 2 (SAM2) demonstrates remarkable universal segmentation capabilities on natural images. However, its performance on ultrasound images is significantly degraded due to domain disparities. This limitation raises two critical challenges: how to efficiently adapt SAM2 to ultrasound imaging while maintaining parameter efficiency, and how to deploy the adapted model effectively in resource-constrained clinical environments. To address these issues, we propose UniUltra for universal ultrasound segmentation. Specifically, we first introduce a novel context-edge hybrid adapter (CH-Adapter) that enhances fine-grained perception across diverse ultrasound imaging modalities while achieving parameter-efficient fine-tuning. To further improve clinical applicability, we develop a deep-supervised knowledge distillation (DSKD) technique that transfers knowledge from the large image encoder of the fine-tuned SAM2 to a super lightweight encoder, substantially reducing computational requirements without compromising performance. Extensive experiments demonstrate that UniUltra outperforms state-of-the-arts with superior generalization capabilities. Notably, our framework achieves competitive performance using only 8.91% of SAM2's parameters during fine-tuning, and the final compressed model reduces the parameter count by 94.08% compared to the original SAM2, making it highly suitable for practical clinical deployment. The source code is available at https://github.com/xq141839/UniUltra.