Existing Two-Dimensional (2D) rectal tumor segmentation methods struggle to fully exploit the relationships between slices due to the absence of Three-Dimensional (3D) spatial information, while traditional 3D segmentation techniques often suffer from poor performance and low rotational robustness. To address these limitations, we propose a 3D Magnetic Resonance (MR) segmentation model based on a deep supervised residual capsule network (namely DRCU-Net). This model introduces a capsule module built on the 3D U-Net architecture, allowing it to capture the spatial hierarchical features of tumor tissue through a dynamic routing mechanism, thereby enhancing the model’s robustness to rotation. Additionally, we employ deep supervision mechanisms to improve model performance and modify the LayerNorm function to extend layer normalization to 3D, facilitating its adaptation to the processing requirements of 3D data. We evaluate our model on the dataset from Shanxi Cancer Hospital (China), it achieves a Dice score of 0.7580 and a Mean Intersection over Union (MIoU) score of 0.6258, which demonstrating its superiority.

Accurate glucose prediction plays an important role in glucose management and closed-loop insulin delivery for subjects with diabetes. Due to its powerful data mining capability, neural networks are used to grasp the glucose trends from continuous glucose monitoring (CGM) data. However, this approach requires a large number of individual data and plentiful computing resources, but has relatively poor interpretability. Given that the glucose metabolism mechanism model contains abundant physiological information, fusing the information with neural networks can reduce the demand for data and computing resources, and improve interpretability. In this study, a physics-informed glucose-insulin neural network (PIGNN) model is proposed, of which the structure and loss function are designed based on the glucose-insulin dynamic model. According to the experiments of 22 real subjects and 30 in-silico subjects with type 1 diabetes, the prediction accuracy of this method achieves 0.726 ± 0.126 mmol/L. Compared with models without physical information, the proposed PIGNN shows a significant improvement in glucose prediction for real subjects with a limited sample size (only 48 data samples), resulting in an accuracy improvement of 12.33%. In addition, it is proved that with limited data more physical information can improve the glucose prediction accuracy significantly.

The hybrid of wireless-powered communication (WPCom) and backscatter communication (BC) is a promising paradigm that will solve the energy shortage in the future Internet of Things (IoT). Adaptive switching between two communication modes can mitigate BC failure in poor channel conditions. However, adaptive switching poses a significant burden on nodes with limited resources. In this paper, we introduce ASNHC, an adaptive switching node that hybridizes BC and WPCom. To address the high power consumption associated with adaptive switching, we designed a low-power signal detection circuit and switching mechanism, enabling the node to operate flexibly across four distinct modes. Furthermore, we leverage the node’s adaptive switching capability to introduce a resource allocation scheme, which coordinates hybrid data transmission to maximize the system sum throughput. Experimental results demonstrate that the node consumes only 2.64μW during mode switching, and the proposed resource allocation scheme significantly enhances throughput compared to the conventional harvest-then-transmit (HTT) protocol.

Using global navigation satellite system (GNSS) to monitor snow depth helps scientists study the impacts of climate change and predict future climate patterns. In the process of extracting reflection signals from signal-to-noise ratio (SNR) data, traditional methods usually use low order polynomials for detrending terms. However, this traditional algorithm cannot completely eliminate the environmental noise in the SNR during signal decomposition, resulting in other noise sources still existing in the detrended SNR, which affects the inversion results. In order to make the inversion results more accurate, this paper proposes a robust empirical mode decomposition (REMD) based method. In signal decomposition, REMD is applied to improve the algorithm, and the correlation coefficient method is used to denoise the decomposed signal. The proposed algorithm is validated using the data from the U.S. Plate Boundary Observatory network SG27 site from winter 2016 to spring 2017 as the study data. The obtained experimental results are compared with the actual snow depth provided by Snowpack Telemetry. When the improved algorithm is used, the root mean square error and mean absolute error of the snow depth inversion at the SG27 site are improved, respectively.

Digital twin technology has emerged as a promising strategy to augment the safety level of intelligent transportation systems by predicting the driving states of neighboring vehicles. Furthermore, with the capacity to anticipate the location of neighboring vehicles, a reduction in driving state exchanges can be achieved, which in turn decreases communication network loads and enhances system performance. However, a theoretical analysis of the performance benefits of digital twin technology in vehicular networks remains a challenge. To address this issue, this paper employs Network Calculus theory to derive the theoretical delay upper bounds of a digital twin-enabled vehicular network. Initially, we analyze the delays of constant interval arrival applications and Poisson arrival applications under Vehicle-to-Vehicle (V2V) communication in the C-V2X Mode 4 protocol. Subsequently, we examine the relationship between the driving state exchange interval and location prediction error within the digital twin framework. These two theoretical models are then integrated to formulate a method for modeling delays under varying tolerance errors. The validity of these theoretical models is confirmed by numerical outcomes. Simulation results indicate that in most scenarios, digital twin technology can diminish network loads, with a typical reduction of approximately 40% in driving state messages. Meanwhile, the average communication delay can be reduced by approximately 10%.

Plant phenotype detection plays a crucial role in understanding and studying plant biology, agriculture, and ecology. It involves the quantification and analysis of various physical traits and characteristics of plants, such as plant height, leaf shape, angle, number, and growth trajectory. By accurately detecting and measuring these phenotypic traits, researchers can gain insights into plant growth, development, stress tolerance, and the influence of environmental factors, which has important implications for crop breeding. Among these phenotypic characteristics, the number of leaves and growth trajectory of the plant are most accessible. Nonetheless, obtaining these phenotypes is labor intensive and financially demanding. With the rapid development of computer vision technology and artificial intelligence, using maize field images to fully analyze plant-related information can greatly eliminate repetitive labor and enhance the efficiency of plant breeding. However, it is still difficult to apply deep learning methods in field environments to determine the number and growth trajectory of leaves and stalks due to the complex backgrounds and serious occlusion problems of crops in field environments. To preliminarily explore the application of deep learning technology to the acquisition of the number of leaves and stalks and the tracking of growth trajectories in field agriculture, in this study, we developed a deep learning method called Point-Line Net, which is based on the Mask R-CNN framework, to automatically recognize maize field RGB images and determine the number and growth trajectory of leaves and stalks. The experimental results demonstrate that the object detection accuracy (mAP50) of our Point-Line Net can reach 81.5%. Moreover, to describe the position and growth of leaves and stalks, we introduced a new lightweight “keypoint” detection branch that achieved a magnitude of 33.5 using our custom distance verification index. Overall, these findings provide valuable insights for future field plant phenotype detection, particularly for datasets with dot and line annotations.

Working as aerial base stations, mobile robotic agents can be formed as a wireless robotic network to provide network services for on-ground mobile devices in a target area. Herein, a challenging issue is how to deploy these mobile robotic agents to provide network services with good quality for more users, while considering the mobility of on-ground devices. In this paper, to solve this issue, we decouple the coverage problem into the vertical dimension and the horizontal dimension without any loss of optimization and introduce the network coverage model with maximum coverage range. Then, we propose a hybrid deployment algorithm based on the improved quick artificial bee colony. The algorithm is composed of a centralized deployment algorithm and a distributed one. The proposed deployment algorithm deploy a given number of mobile robotic agents to provide network services for the on-ground devices that are independent and identically distributed. Simulation results have demonstrated that the proposed algorithm deploys agents appropriately to cover more ground area and provide better coverage uniformity.

Severe cardiovascular diseases can rapidly lead to death. At present, most studies in the deep learning field using electrocardiogram (ECG) are performed on intra-patient experiments for the classification of coronary artery disease (CAD), myocardial infarction, and congestive heart failure (CHF). By contrast, actual conditions are inter-patient experiments. In this study, we proposed a deep learning network, namely, CResFormer, with dual feature extraction to improve accuracy in classifying such diseases. First, fixed segmentation of dual-lead ECG signals without preprocessing was used as input data. Second, one-dimensional convolutional layers performed moderate dimensionality reduction to accommodate subsequent feature extraction. Then, ResNet residual network block layers and transformer encoder layers sequentially performed feature extraction to obtain key associated abstract features. Finally, the Softmax function was used for classifications. Notably, the focal loss function is used when dealing with unbalanced datasets. The average accuracy, sensitivity, positive predictive value, and specificity of four classifications of severe cardiovascular diseases are 99.84%, 99.68%, 99.71%, and 99.90% in intra-patient experiments, respectively, and 97.48%, 93.54%, 96.30%, and 97.89% in inter-patient experiments, respectively. In addition, the model performs well in unbalanced datasets and shows good noise robustness. Therefore, the model has great application potential in diagnosing CAD, MI, and CHF in the actual clinical environment.

Although deep learning methods have recently attracted considerable attention in the medical field, analyzing large-scale electronic health record data is still a difficult task. In particular, the accurate recognition of heart failure is a key technology for doctors to make reasonable treatment decisions. This study uses data from the Medical Information Mart for Intensive Care database. Compared with structured data, unstructured data contain abundant patient information. However, this type of data has unsatisfactory characteristics, e.g., many colloquial vocabularies and sparse content. To solve these problems, we propose the KTI-RNN model for unstructured data recognition. The proposed model overcomes sparse content and obtains good classification results. The term frequency-inverse word frequency (TF-IWF) model is used to extract the keyword set. The latent dirichlet allocation (LDA) model is adopted to extract the topic word set. These models enable the expansion of the medical record text content. Finally, we embed the global attention mechanism and gating mechanism between the bidirectional recurrent neural network (BiRNN) model and the output layer. We call it gated-attention-BiRNN (GA-BiRNN) and use it to identify heart failure from extensive medical texts. Results show that the $F\mathsf{}\mathrm{1}$ score of the proposed KTI-RNN model is 85.57%, and the accuracy rate of the proposed KTI-RNN model is 85.59%.

Computational Radio Frequency IDentification (CRFID) is a device that integrates passive sensing and computing applications, which is powered by electromagnetic waves and read by the off-the-shelf Ultra High Frequency Radio Frequency IDentification (UHF RFID) readers. Traditional RFID only identifies the ID of the tag, and CRFID is different from traditional RFID. CRFID needs to transmit a large amount of sensing and computing data in the mobile sensing scene. However, the current Electronic Product Code, Class-1 Generation-2 (EPC C1G2) protocol mainly aims at the transmission of multi-tag and minor data. When a large amount of data need to be fed back, a more reliable communication mechanism must be used to ensure the efficiency of data exchange. The main strategy of this paper is to adjust the data frame length of the CRFID response dynamically to improve the efficiency and reliability of CRFID backscattering communication according to energy acquisition and channel complexity. This is done by constructing a dynamic data frame length model and optimizing the command set of the interface protocol. Then, according to the actual situation of the uplink, a dynamic data validation method is designed, which reduces the data transmission delay and the probability of retransmitting, and improves the throughput. The simulation results show that the proposed scheme is superior to the existing methods. Under different energy harvesting and channel conditions, the dynamic data frame length and verification method can approach the theoretical optimum.