Sepsis Prediction Using CNNBDLSTM and Temporal Derivatives Feature Extraction in the IoT Medical Environment

Background:Sepsis,a potentially fatal inflammatory disease triggered by infection,carries significant healthimplications worldwide.Timely detection is crucial as sepsis can rapidly escalate if left undetected.Recentadvancements in deep learning(DL)offer powerful tools to address this challenge.Aim:Thus,this study proposeda hybrid CNNBDLSTM,a combination of a convolutional neural network(CNN)with a bi-directional long shorttermmemory(BDLSTM)model to predict sepsis onset.Implementing the proposed model provides a robustframework that capitalizes on the complementary strengths of both architectures,resulting in more accurate andtimelier predictions.Method:The sepsis prediction method proposed here utilizes temporal feature extraction todelineate six distinct time frames before the onset of sepsis.These time frames adhere to the sepsis-3 standardrequirement,which incorporates 12-h observation windows preceding sepsis onset.All models were trained usingthe Medical Information Mart for Intensive Care III(MIMIC-III)dataset,which sourced 61,522 patients with 40clinical variables obtained from the IoT medical environment.The confusion matrix,the area under the receiveroperating characteristic curve(AUCROC)curve,the accuracy,the precision,the F1-score,and the recall weredeployed to evaluate themodels.Result:The CNNBDLSTMmodel demonstrated superior performance comparedto the benchmark and other models,achieving an AUCROC of 99.74%and an accuracy of 99.15%one hour beforesepsis onset.These results indicate that the CNNBDLSTM model is highly effective in predicting sepsis onset,particularly within a close proximity of one hour.Implication:The results could assist practitioners in increasingthe potential survival of the patient one hour before sepsis onset.

Smart Healthcare Activity Recognition Using Statistical Regression and Intelligent Learning

In this present time,Human Activity Recognition(HAR)has been of considerable aid in the case of health monitoring and recovery.The exploitation of machine learning with an intelligent agent in the area of health informatics gathered using HAR augments the decision-making quality and significance.Although many research works conducted on Smart Healthcare Monitoring,there remain a certain number of pitfalls such as time,overhead,and falsification involved during analysis.Therefore,this paper proposes a Statistical Partial Regression and Support Vector Intelligent Agent Learning(SPR-SVIAL)for Smart Healthcare Monitoring.At first,the Statistical Partial Regression Feature Extraction model is used for data preprocessing along with the dimensionality-reduced features extraction process.Here,the input dataset the continuous beat-to-beat heart data,triaxial accelerometer data,and psychological characteristics were acquired from IoT wearable devices.To attain highly accurate Smart Healthcare Monitoring with less time,Partial Least Square helps extract the dimensionality-reduced features.After that,with these resulting features,SVIAL is proposed for Smart Healthcare Monitoring with the help of Machine Learning and Intelligent Agents to minimize both analysis falsification and overhead.Experimental evaluation is carried out for factors such as time,overhead,and false positive rate accuracy concerning several instances.The quantitatively analyzed results indicate the better performance of our proposed SPR-SVIAL method when compared with two state-of-the-art methods.

Image Enhancement Using Adaptive Fractional Order Filter

Image enhancement is an important preprocessing task as the contrast is low in most of the medical images,Therefore,enhancement becomes the mandatory process before actual image processing should start.This research article proposes an enhancement of the model-based differential operator for the images in general and Echocardiographic images,the proposed operators are based on Grunwald-Letnikov(G-L),Riemann-Liouville(R-L)and Caputo(Li&Xie),which are the definitions of fractional order calculus.In this fractional-order,differentiation is well focused on the enhancement of echocardiographic images.This provoked for developing a non-linear filter mask for image enhancement.The designed filter is simple and effective in terms of improving the contrast of the input low contrast images and preserving the textural features,particularly in smooth areas.The novelty of the proposed method involves a procedure of partitioning the image into homogenous regions,details,and edges.Thereafter,a fractional differential mask is appropriately chosen adaptively for enhancing the partitioned pixels present in the image.It is also incorporated into the Hessian matrix with is a second-order derivative for every pixel and the parameters such as average gradient and entropy are used for qualitative analysis.The wide range of existing state-of-the-art techniques such as fixed order fractional differential filter for enhancement,histogram equalization,integer-order differential methods have been used.The proposed algorithm resulted in the enhancement of the input images with an increased value of average gradient as well as entropy in comparison to the previous methods.The values obtained are very close(almost equal to 99.9%)to the original values of the average gradient and entropy of the images.The results of the simulation validate the effectiveness of the proposed algorithm.

Machine Learning Based Diagnosis for Diabetic Retinopathy for SKPD-PSC

The study aimed to apply to Machine Learning(ML)researchers working in image processing and biomedical analysis who play an extensive role in compre-hending and performing on complex medical data,eventually improving patient care.Developing a novel ML algorithm specific to Diabetic Retinopathy(DR)is a chal-lenge and need of the hour.Biomedical images include several challenges,including relevant feature selection,class variations,and robust classification.Although the cur-rent research in DR has yielded favourable results,several research issues need to be explored.There is a requirement to look at novel pre-processing methods to discard irrelevant features,balance the obtained relevant features,and obtain a robust classi-fication.This is performed using the Steerable Kernalized Partial Derivative and Platt Scale Classifier(SKPD-PSC)method.The novelty of this method relies on the appropriate non-linear classification of exclusive image processing models in har-mony with the Platt Scale Classifier(PSC)to improve the accuracy of DR detection.First,a Steerable Filter Kernel Pre-processing(SFKP)model is applied to the Retinal Images(RI)to remove irrelevant and redundant features and extract more meaningful pathological features through Directional Derivatives of Gaussians(DDG).Next,the Partial Derivative Image Localization(PDIL)model is applied to the extracted fea-tures to localize candidate features and suppress the background noise.Finally,a Platt Scale Classifier(PSC)is applied to the localized features for robust classification.For the experiments,we used the publicly available DR detection database provided by Standard Diabetic Retinopathy(SDR),called DIARETDB0.A database of 130 image samples has been collected to train and test the ML-based classifiers.Experimental results show that the proposed method that combines the image processing and ML models can attain good detection performance with a high DR detection accu-racy rate with minimum time and complexity compared to the state-of-the-art meth-ods.The accuracy and speed of DR detection for numerous types of images will be tested through experimental evaluation.Compared to state-of-the-art methods,the method increases DR detection accuracy by 24%and DR detection time by 37.

Optimized Resource Allocation and Queue Management for Traffic Control in MANET

A set of mobile devices that employs wireless transmission for communication is termed Mobile Ad hoc Networks(MANETs).Offering better communication services among the users in a centralized organization is the primary objective of the MANET.Due to the features of MANET,this can directly End-to-End Delay(EED)the Quality of Service(QoS).Hence,the implementation of resource management becomes an essential issue in MANETs.This paper focuses on the efficient Resource Allocation(RA)for many types of Traffic Flows(TF)in MANET.In Mobile Ad hoc Networks environments,the main objective of Resource Allocation(RA)is to process consistently available resources among terminals required to address the service requirements of the users.These three categories improve performance metrics by varying transmission rates and simulation time.For solving that problem,the proposed work is divided into Queue Management(QM),Admission Control(AC)and RA.For effective QM,this paper develops a QM model for elastic(EL)and inelastic(IEL)Traffic Flows.This research paper presents an AC mechanism for multiple TF for effective AC.This work presents a Resource Allocation Using Tokens(RAUT)for various priority TF for effective RA.Here,nodes have three cycles which are:Non-Critical Section(NCS),Entry Section(ES)and Critical Section(CS).When a node requires any resources,it sends Resource Request Message(RRM)to the ES.Elastic and inelastic TF priority is determined using Fuzzy Logic(FL).The token holder selects the node from the inelastic queue with high priority for allocating the resources.Using Network Simulator-2(NS-2),simulations demonstrate that the proposed design increases Packet Delivery Ratio(PDR),decrease Packet Loss Ratio(PLR),minimise the Fairness and reduce the EED.

Improved Multi-Path Routing for QoS on MANET

A Mobile Ad hoc NETwork(MANET)is a self-configuring network that is not reliant on infrastructure.This paper introduces a new multipath routing method based on the Multi-Hop Routing(MHR)technique.MHR is the consecutive selection of suitable relay nodes to send information across nodes that are not within direct range of each other.Failing to ensure good MHR leads to several negative consequences,ultimately causing unsuccessful data transmission in a MANET.This research work consists of three portions.The first to attempt to propose an efficient MHR protocol is the design of Priority Based Dynamic Routing(PBDR)to adapt to the dynamic MANET environment by reducing Node Link Failures(NLF)in the network.This is achieved by dynamically considering a node’s mobility parameters like relative velocity and link duration,which enable the next-hop selection.This method works more efficiently than the traditional protocols.Then the second stage is the Improved Multi-Path Dynamic Routing(IMPDR).The enhancement is mainly focused on further improving the Quality of Service(QoS)in MANETs by introducing a QoS timer at every node to help in the QoS routing of MANETs.Since QoS is the most vital metric that assesses a protocol,its dynamic estimation has improved network performance considerably.This method uses distance,linkability,trust,and QoS as the four parameters for the next-hop selection.IMPDR is compared against traditional routing protocols.The Network Simulator-2(NS2)is used to conduct a simulation analysis of the protocols under consideration.The proposed tests are assessed for the Packet Delivery Ratio(PDR),Packet Loss Rate(PLR),End-to-End Delay(EED),and Network Throughput(NT).