Drone-view object detection models operating under low-light conditions face several challenges, such as object scale variations, high image noise, and limited computational resources. Existing models often struggle to balance accuracy and lightweight architecture. This paper introduces ELS-YOLO, a lightweight object detection model tailored for low-light environments, built upon the YOLOv11s framework. ELS-YOLO features a re-parameterized backbone (ER-HGNetV2) with integrated Re-parameterized Convolution and Efficient Channel Attention mechanisms, a Lightweight Feature Selection Pyramid Network (LFSPN) for multi-scale object detection, and a Shared Convolution Separate Batch Normalization Head (SCSHead) to reduce computational complexity. Layer-Adaptive Magnitude-Based Pruning (LAMP) is employed to compress the model size. Experiments on the ExDark and DroneVehicle datasets demonstrate that ELS-YOLO achieves high detection accuracy with a compact model. Here, we show that ELS-YOLO attains a mAP@0.5 of 74.3% and 68.7% on the ExDark and DroneVehicle datasets, respectively, while maintaining real-time inference capability.
For a large population of elderly who live alone, a continuous long-term daily monitoring system is critical and imminently needed to enhance the quality of their lives. Continuous monitoring of vitality information (i.e., which area the elder is staying in, the motion state and activity intensity of an elder) is essential for elderly care. In this paper, we use existing commodity Wi-Fi devices to design and implement a long-term device-free human daily vitality system, WiMonitor. Our system can continuously capture the target’s vitality information in a multi-room home environment without compromising the privacy of the target. In a continuous 22-day experiment, WiMonitor successfully captures the human vitality information accurately. We believe our system can provide valuable long-term monitoring data for both researchers and health care personnel.
Drone-view object detection models operating under low-light conditions face several challenges, such as object scale variations, high image noise, and limited computational resources. Existing models often struggle to balance accuracy and lightweight architecture. This paper introduces ELS-YOLO, a lightweight object detection model tailored for low-light environments, built upon the YOLOv11s framework. ELS-YOLO features a re-parameterized backbone (ER-HGNetV2) with integrated Re-parameterized Convolution and Efficient Channel Attention mechanisms, a Lightweight Feature Selection Pyramid Network (LFSPN) for multi-scale object detection, and a Shared Convolution Separate Batch Normalization Head (SCSHead) to reduce computational complexity. Layer-Adaptive Magnitude-Based Pruning (LAMP) is employed to compress the model size. Experiments on the ExDark and DroneVehicle datasets demonstrate that ELS-YOLO achieves high detection accuracy with a compact model. Here, we show that ELS-YOLO attains a mAP@0.5 of 74.3% and 68.7% on the ExDark and DroneVehicle datasets, respectively, while maintaining real-time inference capability.
RGB-infrared (RGB-IR) object detection leverages complementary information from these two modalities to substantially enhance perception in complex environments, which is particularly beneficial for reliable detection under adverse imaging conditions such as low illumination and severe haze. However, RGB-IR object detection still faces several challenges due to pronounced intra-modality and cross-modality discrepancies. On the one hand, many existing approaches rely on complex architectures to strengthen cross-modal interactions, which increases computational cost. On the other hand, symmetric dual-branch backbones with a static fusion paradigm often struggle to explicitly characterize discrepancies between the RGB and IR modalities. This limitation prevents effective mining of complementary information and reduces the discriminability of fused representations. To address these issues, this paper presents a lightweight RGB-IR multimodal detection network (LMDENet), which consists of three key components: (1) an illumination-guided label selection (IGLS) that integrates RGB and IR labels based on cross-modal matching and illumination-aware rules to construct consistent and reliable supervision; (2) a heterogeneous backbone network (HBN) with differentiated branches that separately model RGB appearance details and IR structural information, improving modality-specific representation learning; and (3) a difference-complement enhancement module (DCEM) that explicitly decomposes cross-modal features into common and difference components and performs selective enhancement to amplify complementary information while suppressing redundant noise. We systematically evaluate the detection performance of the proposed model on the multimodal remote sensing dataset DroneVehicle, and further conduct supplementary experiments on the LLVIP dataset to verify its generalization ability across different scenarios. Experimental results on the DroneVehicle and LLVIP datasets demonstrate that LMDENet achieves and mAP@0.5, respectively. Meanwhile, the model contains only 3.3 M parameters and 8.7 G FLOPs, reflecting a favorable accuracy–efficiency balance.
Nowadays, Liquid identification plays an important role in our daily lives, especially for security checks at railway stations or airports. Traditional identification technologies are ether intrusive or expensive, so they are inconvenient for use. In this paper, we use commercial off-the-shelf (COTS) RFID devices to design and implement a non-intrusive, contactless liquid identification system, RF-Symmetry. Based on the symmetrical characteristics of refraction when RF signals travel through the cylindrical container, we use two symmetrically-located RFID Tag Arrays to receive RF signals which are transmitted by RFID antenna. In order to effectively select the symmetrical tag pairs, we design a tag selection algorithm to select tags from two tag arrays. We design a specialized signal processing method to extract liquid features, and then use extreme learning machine optimized by genetic algorithm (GA-ELM) to classify liquids. We build a prototype system of RF-Symmetry, through extensive experiments, we can find that RF-Symmetry can achieve a high recognition accuracy (97.22%) on 12 different liquids. Moreover, the experiments also show that our system can effectively distinguish similar liquids such as coke and diet coke.
The growing demand for personalized healthcare is driving the development of Wireless Body Area Networks (WBANs). These networks enable continuous monitoring of physiological parameters. In WBANs, routing protocols are essential for ensuring reliable data delivery. However, designing efficient protocols is challenging due to the specific environment of the human body. Key issues include limited energy, frequent topology changes caused by movement, and diverse Quality of Service needs. In this review, we investigate, summarize, and analyze state-of-the-art WBAN routing protocols. Specifically, we outline the architecture of WBAN-based eHealth systems and review major design challenges. We then present a categorized survey of recent protocols. Subsequently, we examine the distribution across protocol categories and compare their performance. Finally, we identify open challenges and discuss future research directions.