Rupsa Rani Mishra, D. Chandrasekhar Rao, Ajaya Kumar Tripathy
Manual observation and monitoring of individual cows for disease detection present significant challenges in large-scale farming operations, as the process is labor-intensive, time-consuming, and prone to reduced accuracy. The reliance on human observation often leads to delays in identifying symptoms, as the sheer number of animals can hinder timely attention to each cow. Consequently, the accuracy and precision of disease detection are significantly compromised, potentially affecting animal health and overall farm productivity. Furthermore, organizing and managing human resources for the manual observation and monitoring of cow health is a complex and economically demanding task. It necessitates the involvement of skilled personnel, thereby contributing to elevated farm maintenance costs and operational inefficiencies. Therefore, the development of an automated, low-cost, and reliable smart system is essential to address these challenges effectively. Although several studies have been conducted in this domain, very few have simultaneously considered the detection of multiple common diseases with high prediction accuracy. However, advancements in Internet of Things (IoT), Machine Learning (ML), and Cyber-Physical Systems have enabled the automation of cow health monitoring with enhanced accuracy and reduced operational costs. This study proposes an IoT-enabled Cyber-Physical System framework designed to monitor the daily activities and health status of cow. A novel ML algorithm is proposed for the diagnosis of common cow diseases using collected physiological and behavioral data. The algorithm is designed to predict multiple diseases by analyzing a comprehensive set of recorded physiological and behavioral features, enabling accurate and efficient health assessment.
We propose an epidemic model for the spread of vector-borne diseases. The model, which is built extending the classical susceptible-infected-susceptible model, accounts for two populations -- humans and vectors -- and for cross-contagion between the two species, whereby humans become infected upon interaction with carrier vectors, and vectors become carriers after interaction with infected humans. We formulate the model as a system of ordinary differential equations and leverage monotone systems theory to rigorously characterize the epidemic dynamics. Specifically, we characterize the global asymptotic behavior of the disease, determining conditions for quick eradication of the disease (i.e., for which all trajectories converge to a disease-free equilibrium), or convergence to a (unique) endemic equilibrium. Then, we incorporate two control actions: namely, vector control and incentives to adopt protection measures. Using the derived mathematical tools, we assess the impact of these two control actions and determine the optimal control policy.
Disease progression modeling aims to characterize and predict how a patient's disease complications worsen over time based on longitudinal electronic health records (EHRs). For diseases such as type 2 diabetes, accurate progression modeling can enhance patient sub-phenotyping and inform effective and timely interventions. However, the problem is challenging due to the need to learn continuous-time progression dynamics from irregularly sampled clinical events amid patient heterogeneity (e.g., different progression rates and pathways). Existing mechanistic and data-driven methods either lack adaptability to learn from real-world data or fail to capture complex continuous-time dynamics on progression trajectories. To address these limitations, we propose Temporally Detailed Hypergraph Neural Ordinary Differential Equation (TD-HNODE), which represents disease progression on clinically recognized trajectories as a temporally detailed hypergraph and learns the continuous-time progression dynamics via a neural ODE framework. TD-HNODE contains a learnable TD-Hypergraph Laplacian that captures the interdependency of disease complication markers within both intra- and inter-progression trajectories. Experiments on two real-world clinical datasets demonstrate that TD-HNODE outperforms multiple baselines in modeling the progression of type 2 diabetes and related cardiovascular diseases.
Anatole Besarab, Stanley Frinak, Suresh Margassery
et al.
This review describes the history of vascular access for hemodialysis (HD) over the past 8 decades. Reliable, repeatable vascular access for outpatient HD began in the 1960s with the Quinton-Scribner shunt. This was followed by the autologous Brecia-Cimino radial-cephalic arteriovenous fistula (AVF), which dominated HD vascular access for the next 20 years. Delayed referral and the requirement of 1.5-3 months for AVF maturation led to the development of and increasing dependence on synthetic arteriovenous grafts (AVGs) and tunneled central venous catheters, both of which have higher thrombosis and infection risks than AVFs. The use of AVGs and tunneled central venous catheters increased progressively to the point that, in 1997, the first evidence-based clinical practice guidelines for HD vascular access recommended that they only be used if a functioning AVF could not be established. Efforts to promote AVF use in the United States during the past 2 decades doubled their prevalence; however, recent practice guidelines acknowledge that not all patients receiving HD are ideally suited for an AVF. Nonetheless, improved referral for AVF placement before dialysis initiation and improved conversion of failing AVGs to AVFs may increase AVF use among patients in whom they are appropriate.
Introduction: Gaining access to the kidney is crucial step in percutaneous nephrolithotomy (PCNL); it has a steep learning curve.
Objective: Describe the mathematical method to predict renal puncture angle and distance based on preoperative computed tomography (CT) measurements. Then evaluating how it correlates with measured values.
Patients and Methods: The study was prospectively designed. After ethical committee approval, the study uses data from preoperative CT to construct a triangle so we can estimate puncture depth and angle. A triangle of three points, the first is point of entry to the pelvicalyceal system (PCS), the second is point on the skin perpendicular to it, and the third where the needle punctures the skin. The needle travel is estimated using the Pythagorean theorem and puncture angle using the inverse sine function. We evaluated 40 punctures in 36 PCNL procedures. After PCS puncture using fluoroscopy-guided triangulation, we measured the needle travel distance and angle to the horizontal plane. Then compared the results with mathematically estimated values.
Results: We targeted posterior lower calyx in 21 (70%) case. The correlation between measured and estimated needle travel distance with Rho coefficient of 0.76 with P < 0.001. The mean difference between the estimated and the measured needle travel was – 0.37 ± 1.2 cm (−2.6–1.6). Measured and estimated angle correlate with Rho coefficient of 0.77 and P < 0.001. The mean difference between the estimated and the measured angle was 2° ± 8° (−21°–16°).
Conclusion: Mathematical estimation of needle depth and angle for gaining access to the kidney correlates well with measured values.
M. M. Khasanov, U. A. Abdufattaev, A. A. Nomanov
et al.
Introduction. Urolithiasis (UL) is a common disease observed in a huge number of people around the world. Supravesical obstruction (SVO) is a less frequent but potentially dangerous phenomenon that occurs when the urinary tract is blocked above the bladder, often resulting in impaired urinary passage and requiring surgery in most cases.Objective. To determine the risk factors of complicated supravesical obstruction in patients with urolithiasis.Materials & methods. We have conducted a retrospective study of diagnostic and treatment results of patients with SVO suffering from UL. Inclusion criteria: patients with established UL and SVO. Exclusion criteria: cancer and/or specific infection leading to SVO. From 2017 to 2019, 6250 patients with SVO were screened and treated at our centre. After assessing compliance with the inclusion and exclusion criteria, 1106 patients were selected for the study. We studied the risk factors leading to complicated course of SVO.Results. We evaluated the risk factors for the development of complicated SVO in patients with UL. All results are presented by EXP(B), with a 95% confidence interval in square brackets: complex kidney stones — 5,326 [2,247 – 9,296], ureteral stones — 12,251 [7,256 – 21,226], double-sided stones — 7,256 [2,158 – 9,255], disease length — 4,324 [1,782 – 8,256], urinary tract infection — 19,258 [4,258 – 26,248], blood leukocytosis — 15,116 [3,985 – 21,256], high serum creatinine — 10,244 [5,269 – 16,254], high serum glucose — 5,226 [3,145 – 11,254].Conclusion. The results of the study suggest that blocking renal and ureteral stones, double-side stones, upper urinary tract infection, low creatinine clearance, diabetes mellitus and disease length are significant risk factors for complicated course of SVO.
As road accident cases are increasing due to the inattention of the driver, automated driver monitoring systems (DMS) have gained an increase in acceptance. In this report, we present a real-time DMS system that runs on a hardware-accelerator-based edge device. The system consists of an InfraRed camera to record the driver footage and an edge device to process the data. To successfully port the deep learning models to run on the edge device taking full advantage of the hardware accelerators, model surgery was performed. The final DMS system achieves 63 frames per second (FPS) on the TI-TDA4VM edge device.
This paper describes the NPU-MSXF system for the IWSLT 2023 speech-to-speech translation (S2ST) task which aims to translate from English speech of multi-source to Chinese speech. The system is built in a cascaded manner consisting of automatic speech recognition (ASR), machine translation (MT), and text-to-speech (TTS). We make tremendous efforts to handle the challenging multi-source input. Specifically, to improve the robustness to multi-source speech input, we adopt various data augmentation strategies and a ROVER-based score fusion on multiple ASR model outputs. To better handle the noisy ASR transcripts, we introduce a three-stage fine-tuning strategy to improve translation accuracy. Finally, we build a TTS model with high naturalness and sound quality, which leverages a two-stage framework, using network bottleneck features as a robust intermediate representation for speaker timbre and linguistic content disentanglement. Based on the two-stage framework, pre-trained speaker embedding is leveraged as a condition to transfer the speaker timbre in the source English speech to the translated Chinese speech. Experimental results show that our system has high translation accuracy, speech naturalness, sound quality, and speaker similarity. Moreover, it shows good robustness to multi-source data.
We numerically study disease dynamics that lead to the disease switching from one host species to another, resulting in diseases gaining the ability to infect, e.g., humans. Unlike previous studies that focused on branching processes starting with the first infected humans, we begin by considering a disease pathogen that initially cannot infect humans. We model the entire process, starting from an infection in the animal population, including mutations that eventually enable the disease to cause an epidemic outbreak in the human population. We use an SIR model on a network consisting of 132 dog and 1320 human nodes, with a single parameter representing the gene of the pathogen. We use numerical large-deviation techniques, specifically the $1/t$ Wang-Landau algorithm, to calculate the potentially very small probability of the host switching event. With this approach we are able to resolve probabilities as small as $10^{-120}$. Additionally the $1/t$ Wang-Landau algorithm allows us to obtain the complete probability density function $P(C)$ of the cumulative fraction $C$ of infected humans, which is an indicator for the severity of the disease in the human population. We also calculate correlations of $C$ with selected quantities $q$ that characterize the outbreak. Due to the application of the rare-event algorithm, this is possible for the entire range of $C$ values.
Yasemin Coşkun Yavuz, Nihal Cetin, Esma Menevşe
et al.
The study aimed to investigate the role of magnesium sulfate prophylaxis in nephrotoxicity caused by colistin. Thirty Wistar Albino rats were divided into four groups: control, colistin, magnesium (Mg), and Mg + colistin. The drugs were administered to the groups for seven days. Urea-creatinine values were measured at the beginning (T0) and end (T1) of the study. Malondialdehyde (MDA) levels were measured in plasma and kidney tissue, glutathione (GSH) levels were analyzed in the erythrocyte and kidney tissues. At the end of the study, the semiquantitative score (SQS) was calculated by the histopathological examination of the kidneys. Urea values significantly decreased in Mg and Mg + colistin groups compared to the baseline (p = 0.013 and p = 0.001). At the time of T1, these groups had significantly lower urea values than the colistin and control groups. Creatinine value was significantly increased in the colistin group compared to baseline (p = 0.005), the creatinine value in the colistin group was significantly higher than the Mg + colistin group (p = 0.011). Plasma MDA levels were significantly higher in the colistin group compared to the other groups at the time of T1 (p < 0.001). The Mg + colistin group had lower renal MDA levels than the colistin group. The colistin group had significantly higher renal tubular grade (p = 0.035), renal affected area (p < 0.001), and SQS (p = 0.001) than the Mg + colistin group. The results of the study suggested that Mg sulfate may have a nephrotoxicity-reducing effect on colistin. Resumen: El objetivo del estudio fue investigar la función de la profilaxis con sulfato de magnesio en la nefrotoxicidad causada por la colistina. Se dividieron 30 ratas Wistar albinas en 4 grupos: control, colistina, magnesio (Mg) y Mg + colistina. Los fármacos se administraron a los grupos durante 7 días. Los valores de urea-creatinina se midieron al principio (T0) y al final (T1) del estudio. Se midieron los niveles de malondialdehído (MDA) en el plasma y el tejido renal, y se analizaron los niveles de glutatión (GSH) en los eritrocitos y el tejido renal. Al final del estudio, se calculó la puntuación semicuantitativa (semiquantitative score [SQS]) mediante el examen histopatológico de los riñones. Los valores de urea disminuyeron significativamente en los grupos de Mg y Mg + colistina en comparación con los valores iniciales (p = 0,013 y p = 0,001). En el momento del T1, estos grupos tenían valores de urea significativamente más bajos que los grupos de colistina y de control. El valor de creatinina se incrementó significativamente en el grupo de colistina en comparación con el valor inicial (p = 0,005); el valor de creatinina en el grupo de colistina fue significativamente mayor que en el grupo de Mg + colistina (p = 0,011). Los niveles de MDA en el plasma fueron significativamente más altos en el grupo de colistina en comparación con los otros grupos en el momento del T1 (p < 0,001). El grupo de Mg + colistina presentó niveles renales de MDA más bajos que el grupo de colistina. El grupo de colistina presentó un grado tubular renal (p = 0,035), un área renal afectada (p < 0,001) y una SQS (p = 0,001) significativamente mayores que el grupo de Mg + colistina. Los resultados del estudio indicaron que el sulfato de Mg puede tener un efecto reductor de la nefrotoxicidad de la colistina.
User queries for a real-world dialog system may sometimes fall outside the scope of the system's capabilities, but appropriate system responses will enable smooth processing throughout the human-computer interaction. This paper is concerned with the user's intent, and focuses on out-of-scope intent classification in dialog systems. Although user intents are highly correlated with the application domain, few studies have exploited such correlations for intent classification. Rather than developing a two-stage approach that first classifies the domain and then the intent, we propose a hierarchical multi-task learning approach based on a joint model to classify domain and intent simultaneously. Novelties in the proposed approach include: (1) sharing supervised out-of-scope signals in joint modeling of domain and intent classification to replace a two-stage pipeline; and (2) introducing a hierarchical model that learns the intent and domain representations in the higher and lower layers respectively. Experiments show that the model outperforms existing methods in terms of accuracy, out-of-scope recall and F1. Additionally, threshold-based post-processing further improves performance by balancing precision and recall in intent classification.
Abu Quwsar Ohi, M. F. Mridha, Farisa Benta Safir
et al.
Clustering is widely used in unsupervised learning method that deals with unlabeled data. Deep clustering has become a popular study area that relates clustering with Deep Neural Network (DNN) architecture. Deep clustering method downsamples high dimensional data, which may also relate clustering loss. Deep clustering is also introduced in semi-supervised learning (SSL). Most SSL methods depend on pairwise constraint information, which is a matrix containing knowledge if data pairs can be in the same cluster or not. This paper introduces a novel embedding system named AutoEmbedder, that downsamples higher dimensional data to clusterable embedding points. To the best of our knowledge, this is the first research endeavor that relates to traditional classifier DNN architecture with a pairwise loss reduction technique. The training process is semi-supervised and uses Siamese network architecture to compute pairwise constraint loss in the feature learning phase. The AutoEmbedder outperforms most of the existing DNN based semi-supervised methods tested on famous datasets.
This paper presents a robust deep learning framework developed to detect respiratory diseases from recordings of respiratory sounds. The complete detection process firstly involves front end feature extraction where recordings are transformed into spectrograms that convey both spectral and temporal information. Then a back-end deep learning model classifies the features into classes of respiratory disease or anomaly. Experiments, conducted over the ICBHI benchmark dataset of respiratory sounds, evaluate the ability of the framework to classify sounds. Two main contributions are made in this paper. Firstly, we provide an extensive analysis of how factors such as respiratory cycle length, time resolution, and network architecture, affect final prediction accuracy. Secondly, a novel deep learning based framework is proposed for detection of respiratory diseases and shown to perform extremely well compared to state of the art methods.
V. Rajinikanth, Nilanjan Dey, Alex Noel Joseph Raj
et al.
Pneumonia is one of the foremost lung diseases and untreated pneumonia will lead to serious threats for all age groups. The proposed work aims to extract and evaluate the Coronavirus disease (COVID-19) caused pneumonia infection in lung using CT scans. We propose an image-assisted system to extract COVID-19 infected sections from lung CT scans (coronal view). It includes following steps: (i) Threshold filter to extract the lung region by eliminating possible artifacts; (ii) Image enhancement using Harmony-Search-Optimization and Otsu thresholding; (iii) Image segmentation to extract infected region(s); and (iv) Region-of-interest (ROI) extraction (features) from binary image to compute level of severity. The features that are extracted from ROI are then employed to identify the pixel ratio between the lung and infection sections to identify infection level of severity. The primary objective of the tool is to assist the pulmonologist not only to detect but also to help plan treatment process. As a consequence, for mass screening processing, it will help prevent diagnostic burden.