Search

Patients’ safety is paramount in the healthcare industry, and reducing medication errors is essential for improvement. A promising solution to this problem involves the development of automated systems capable of assisting patients in verifying their pill intake mistakes. This paper investigates a Pill-Prescription matching task that seeks to associate pills in a multi-pill photo with their corresponding names in the prescription. We specifically aim to overcome the limitations of existing pill detection methods when faced with unseen pills, a situation characteristic of zero-shot learning. We propose a novel method named Zero-PIMA (Zero-shot Pill-Prescription Matching), designed to match pill images with prescription names effectively, even for pills not included in the training dataset. Zero-PIMA is an end-to-end model that includes an object localization module to determine and extract features of pill images and a graph convolutional network to capture the spatial relationship of the pills’ text in the prescription. After that, we leverage the contrastive learning paradigm to increase the distance between mismatched pill images and pill name pairs while minimizing the distance between matched pairs. In addition, to deal with the zero-shot pill detection problem, we leverage pills’ metadata retrieved from the DrugBank database to fine-tune a pre-trained text encoder, thereby incorporating visual information about pills (e.g., shape, color) into their names, making them more informative and ultimately enhancing the pill image-name matching accuracy. Extensive experiments are conducted on our collected real-world VAIPEPP dataset of multi-pill photos and prescriptions. Through a series of comprehensive experiments, the proposed method outperforms other methods for both seen and unseen pills in terms of mean average precision. These results indicate that the proposed method could reduce medication errors and improve patients’ safety.

Abstract Forecasting discharge (Q) and water level (H) are essential factors in hydrological research and flood prediction. In recent years, deep learning has emerged as a viable technique for capturing the non-linear relationship of historical data to generate highly accurate prediction results. Despite the success in various domains, applying deep learning in Q and H prediction is hampered by three critical issues: a shortage of training data, the occurrence of noise in the collected data, and the difficulty in adjusting the model’s hyper-parameters. This work proposes a novel deep learning-based Q–H prediction model that overcomes all the shortcomings encountered by existing approaches. Specifically, to address data scarcity and increase prediction accuracy, we design an ensemble learning architecture that takes advantage of multiple deep learning techniques. Furthermore, we leverage the Singular-Spectrum Analysis (SSA) to remove noise and outliers from the original data. Besides, we exploit the Genetic Algorithm (GA) to propose a novel mechanism that can automatically determine the prediction model’s optimal hyper-parameters. We conducted extensive experiments on two datasets collected from Vietnam’s Red and Dakbla rivers. The results show that our proposed solution outperforms current techniques across a wide range of metrics, including NSE, MSE, MAE, and MAPE. Specifically, by exploiting the ensemble learning technique, we can improve the NSE by at least $$2\%$$ 2 % . Moreover, with the aid of the SSA-based data preprocessing technique, the NSE is further enhanced by more than $$5\%$$ 5 % . Finally, thanks to GA-based optimization, our proposed model increases the NSE by at least $$6\%$$ 6 % and up to $$40\%$$ 40 % in the best case.

Multipath communication is a well-developed technology that enhances communication effectiveness and resilience. Moreover, it can flexibly utilize network resources through load balancing among available paths. However, traditionally, deploying such load balancing functions on network devices is costly due to the required configuration changes and complicated signaling mechanisms on the devices’ control planes. Programming Protocol-independent Packet Processors (P4) has recently emerged as a programming language that enables programmability on the data plane, with the potential to relieve such issues in multipath communication. This work introduces and implements three P4-based multipath schedulers that can split traffic over several paths in wireless networks. The first is P4-based Random Splitting, which distributes traffic randomly. The second is P4-based Weighted Round Robin, with path scheduling based on weights in accordance with path capability. The last is P4-based Dynamic Weighted Round Robin (DWRR), which can improve bandwidth utilization by shifting the weights following dynamic changes in the available bandwidth (i.e., when congestion occurs). We have extensively evaluated the implementation of these three P4-based schedulers in a Mininet-WiFi/P4 environment with User Datagram Protocol (UDP) traffic. The results show that these schedulers can achieve multipath communication with the designed scheduling mechanisms.

Due to the significant resemblance in visual appearance, pill misuse is prevalent and has become a critical issue, responsible for one-third of all deaths worldwide. Pill identification, thus, is a crucial concern that needs to be investigated thoroughly. Recently, several attempts have been made to exploit deep learning to tackle the pill identification problem. However, most published works consider only single-pill identification and fail to distinguish hard samples with identical appearances. Also, most existing pill image datasets only feature single pill images captured in carefully controlled environments under ideal lighting conditions and clean backgrounds. In this work, we are the first to tackle the multi-pill detection problem in real-world settings, aiming at localizing and identifying pills captured by users during pill intake. Moreover, we also introduce a multi-pill image dataset taken in unconstrained conditions. To handle hard samples, we propose a novel method for constructing heterogeneous a priori graphs incorporating three forms of inter-pill relationships, including co-occurrence likelihood, relative size, and visual semantic correlation. We then offer a framework for integrating a priori with pills' visual features to enhance detection accuracy. Our experimental results have proved the robustness, reliability, and explainability of the proposed framework. Experimentally, it outperforms all detection benchmarks in terms of all evaluation metrics. Specifically, our proposed framework improves COCO mAP metrics by 9.4% over Faster R-CNN and 12.0% compared to vanilla YOLOv5. Our study opens up new opportunities for protecting patients from medication errors using an AI-based pill identification solution.

The concentration of fine particulate matter (PM2.5), which represents inhalable particles with diameters of 2.5 micrometers and smaller, is a vital air quality index. Such particles can penetrate deep into the human lungs and severely affect human health. This paper studies accurate PM2.5 prediction, which can potentially contribute to reducing or avoiding the negative consequences. Our approach’s novelty is to utilize the genetic algorithm (GA) and an encoder-decoder (E-D) model for PM2.5 prediction. The GA benefits feature selection and remove outliers to enhance the prediction accuracy. The encoder-decoder model with long short-term memory (LSTM), which relaxes the restrictions between the input and output of the model, can be used to effectively predict the PM2.5 concentration. We evaluate the proposed model on air quality datasets from Hanoi and Taiwan. The evaluation results show that our model achieves excellent performance. By merely using the E-D model, we can obtain more accurate (up to 53.7%) predictions than those of previous works. Moreover, the GA in our model has the advantage of obtaining the optimal feature combination for predicting the PM2.5 concentration. By combining the GA-based feature selection algorithm and the E-D model, our proposed approach further improves the accuracy by at least 13.7%.

This paper addresses the local minimum phenomenon, routing path enlargement, and load imbalance problems of geographic routing in wireless sensor networks (WSNs) with holes. These issues may degrade the network lifetime of WSNs since they cause a long detour path and a traffic concentration around the hole boundary. Aiming to solve these problems, in this work, we propose a novel geographic routing protocol for WSNs, namely, Q-learning Inspired Hole bypassing (QIH), which is lightweight and efficient. QIH’s conceptual idea is to leverage Q-learning to estimate the distance from a node to the holes. QIH makes routing decisions following the nodes’ residual energy, their estimated distance to the holes, and their distance to the destination. We first confirm the effectiveness of QIH by theoretical analysis. Then, we conduct extensive simulations of QIH in comparison to state-of-the-art protocols. The simulation results show that QIH outperforms the other protocols in terms of network lifetime, packet latency, and energy consumption.

In wireless rechargeable sensor networks (WRSNs), a mobile charger (MC) moves around to compensate for sensor nodes’ energy via a wireless medium. In such a context, designing a charging strategy that optimally prolongs the network lifetime is challenging. This work aims to solve the challenges by introducing a novel, on-demand charging algorithm for MC that attempts to maximize the network lifetime, where the term “network lifetime” is defined by the interval from when the network starts till the first target is not monitored by any sensor. The algorithm, named Fuzzy Q-charging, optimizes both the time and location in which the MC performs its charging tasks. Fuzzy Q-charging uses Fuzzy logic to determine the optimal charging-energy amounts for sensors. From that, we propose a method to find the optimal charging time at each charging location. Fuzzy Q-charging leverages Q-learning to determine the next charging location for maximizing the network lifetime. To this end, Q-charging prioritizes the sensor nodes following their roles and selects a suitable charging location where MC provides sufficient power for the prioritized sensors. We have extensively evaluated the effectiveness of Fuzzy Q-charging in comparison to the related works. The evaluation results show that Fuzzy Q-charging outperforms the others. First, Fuzzy Q-charging can guarantee an infinite lifetime in the WSRNs, which have a sufficient large sensor number or a commensurate target number. Second, in other cases, Fuzzy Q-charging can extend the time until the first target is not monitored by 6.8 times on average and 33.9 times in the best case, compared to existing algorithms.

In wireless sensor networks (WSNs) with holes, designing efficient routing protocols, which prolong the network lifetime, is one of the most critical issues. To this end, this paper proposes a new geographic routing protocol for WSNs named the load Balanced and constant Stretch protocol for bypassing Multiple Holes (i.e., BSMH). In contrast to the existing works in the literature, the design of BSMH simultaneously takes into account the three factors that impacted the network lifetime, including routing path length, control packet overhead, and load balance among the nodes. Moreover, BSMH aims at minimizing the routing path length and the control overhead, while maximizing the load balance. We theoretically prove the efficiency of BSMH and extensively evaluate BSMH against the state-of-the-art protocols. The evaluation results show that the proposed protocol outperforms the others in various investigated metrics, not only network lifetime, but also routing path stretch, load balance, and control overhead. Specifically, BSMH prolongs the network lifetime by 30 % compared to the existing protocols while guaranteeing that the routing path stretch is under 1.3 .