Knowledge distillation model for Acute Lymphoblastic Leukemia Detection: Exploring the impact of nesterov-accelerated adaptive moment estimation optimizer

Biomedical Signal Processing and Control, Год журнала: 2024, Номер 94, С. 106246 - 106246

Опубликована: Март 30, 2024

Язык: Английский

---

Design and development of artificial intelligence‐based application programming interface for early detection and diagnosis of colorectal cancer from wireless capsule endoscopy images

International Journal of Imaging Systems and Technology, Год журнала: 2024, Номер 34(2)

Опубликована: Фев. 5, 2024

Abstract The colorectal cancer (CRC) is gaining attention in the context of gastrointestinal tract diseases as it ranks third among most prevalent type cancer. early diagnosis CRC can be done by periodic examination colon and rectum for innocuous tissue abnormality called polyp has potential to evolve malignant future. using wireless capsule endoscopy requires dedicated commitment medical expert demanding significant time, focus effort. accuracy manual analysis identifying polyps extensively reliant on cognitive condition physician, thus emphasizing requirement automatic identification. artificial intelligence integrated computer‐aided system could assist clinician better diagnosis, thereby reducing miss‐rates polyps. In our proposed study, we developed an application program interface aid segmentation evaluation its dimension placement four landmarks predicted polyp. performed light weight Padded U‐Net effective images. We trained validated with augmented images Kvasir dataset calculated performance parameters. order facilitate image augmentation, a graphical user Augment Tree was developed, which incorporates 92 augmentation techniques. accuracy, recall, precision, IoU, F1‐score, loss achieved during validation were 95.6%, 0.946%, 0.985%, 0.933%, 0.965% 0.080% respectively. demonstrated that improved reduced when model rather than only limited original On comparison U‐net architecture recently architectures, attained optimal all metrics except marginal highest value.

Язык: Английский

---

Deep learning-driven prediction in healthcare systems: Applying advanced CNNs for enhanced breast cancer detection

Computers in Biology and Medicine, Год журнала: 2025, Номер 189, С. 109858 - 109858

Опубликована: Фев. 27, 2025

Язык: Английский

---

A Novel IDS with a Dynamic Access Control Algorithm to Detect and Defend Intrusion at IoT Nodes

Sensors, Год журнала: 2024, Номер 24(7), С. 2188 - 2188

Опубликована: Март 29, 2024

The Internet of Things (IoT) is the underlying technology that has enabled connecting daily apparatus to and enjoying facilities smart services. IoT marketing experiencing an impressive 16.7% growth rate a nearly USD 300.3 billion market. These eye-catching figures have made it attractive playground for cybercriminals. devices are built using resource-constrained architecture offer compact sizes competitive prices. As result, integrating sophisticated cybersecurity features beyond scope computational capabilities IoT. All these contributed surge in intrusion. This paper presents LSTM-based Intrusion Detection System (IDS) with Dynamic Access Control (DAC) algorithm not only detects but also defends against novel approach achieved 97.16% validation accuracy. Unlike most IDSs, model proposed IDS been selected optimized through mathematical analysis. Additionally, boasts ability identify wider range threats (14 be exact) compared other solutions, translating enhanced security. Furthermore, fine-tuned strike balance between accurately flagging minimizing false alarms. Its performance metrics (precision, recall, F1 score all hovering around 97%) showcase potential this innovative elevate detection rate, exceeding 98%. high accuracy instills confidence its reliability. lightning-fast response time, averaging under 1.2 s, positions among fastest intrusion systems available.

Язык: Английский

---

Deep-IDS: A Real-Time Intrusion Detector for IoT Nodes Using Deep Learning

IEEE Access, Год журнала: 2024, Номер 12, С. 63584 - 63597

Опубликована: Янв. 1, 2024

The Internet of Things (IoT) represents a swiftly expanding sector that is pivotal in driving the innovation today's smart services. However, inherent resource-constrained nature IoT nodes poses significant challenges embedding advanced algorithms for cybersecurity, leading to an escalation cyberattacks against these nodes. Contemporary research Intrusion Detection Systems (IDS) predominantly focuses on enhancing IDS performance through sophisticated algorithms, often overlooking their practical applicability. This paper introduces Deep-IDS, innovative and practically deployable Deep Learning (DL)-based IDS. It employs Long-Short-Term-Memory (LSTM) network comprising 64 LSTM units trained CIC-IDS2017 dataset. Its streamlined architecture renders Deep-IDS ideal candidate edge-server deployment, acting as guardian between Denial Service (DoS), Distributed (DDoS), Brute Force (BRF), Man-in-the-Middle (MITM), Replay (RP) Attacks. A distinctive aspect this trade-off analysis intrusion detection rate false alarm rate, facilitating real-time Deep-IDS. system demonstrates exemplary 96.8% overall classification accuracy 97.67%. Furthermore, achieves precision, recall, F1-scores 97.67%, 98.17%, 97.91%, respectively. On average, requires 1.49 seconds identify mitigate attempts, effectively blocking malicious traffic sources. remarkable efficacy, swift response time, design, novel defense strategy not only secure but also interconnected sub-networks, thereby positioning IoT-enhanced computer networks.

Язык: Английский

---

Machine learning integrated graphene oxide‐based diagnostics, drug delivery, analytical approaches to empower cancer diagnosis

BMEMat, Год журнала: 2024, Номер unknown

Опубликована: Сен. 28, 2024

Abstract Machine learning (ML) and nanotechnology interfacing are exploring opportunities for cancer treatment strategies. To improve therapy, this article investigates the synergistic combination of Graphene Oxide (GO)‐based devices with ML techniques. The production techniques functionalization tactics used to modify physicochemical characteristics GO specific drug delivery explained at outset investigation. is a great option treating because its natural biocompatibility capacity absorb medicinal chemicals. Then, complicated biological data analyzed using algorithms, which make it possible identify best medicine formulations individualized plans depending on each patient's particular characteristics. study also looks optimizing predicting interactions between carriers cells ML. Predictive modeling helps ensure effective payload release therapeutic efficacy in design customized systems. Furthermore, tracking outcomes real time made by permit adaptive modifications therapy regimens. By medication doses settings, not only decreases adverse effects but enhances accuracy.

Язык: Английский

---

Prediction of Tuberculosis From Lung Tissue Images of Diversity Outbred Mice Using Jump Knowledge Based Cell Graph Neural Network

IEEE Access, Год журнала: 2024, Номер 12, С. 17164 - 17194

Опубликована: Янв. 1, 2024

Tuberculosis (TB), primarily affecting the lungs, is caused by bacterium Mycobacterium tuberculosis and poses a significant health risk. Detecting acid-fast bacilli (AFB) in stained samples critical for TB diagnosis. Whole Slide (WS) Imaging allows digitally examining these samples. However, current deep-learning approaches to analyzing large-sized whole slide images (WSIs) often employ patch-wise analysis, potentially missing complex spatial patterns observed granuloma essential accurate classification. To address this limitation, we propose an approach that models cell characteristics interactions as graph, capturing both cell-level information overall tissue micro-architecture. This method differs from strategies related graph-based works rely on edge thresholds based sparsity/density graph construction, emphasizing biologically informed threshold determination instead. We introduce jumping knowledge neural network (CG-JKNN) operates graphs where are selected length of mycobacteria's cords activated macrophage nucleus's size reflect actual biological tissue. The primary process involves training Convolutional Neural Network (CNN) segment AFBs nuclei, followed converting large (42831*41159 pixels) lung histology into nucleus/AFB represents each node within their denoted edges. enhance interpretability our model, Integrated Gradients Shapely Additive Explanations (SHAP). Our analysis incorporated combination 33 metrics 20 morphology features. In terms traditional machine learning models, Extreme Gradient Boosting (XGBoost) was best performer, achieving F1 score 0.9813 Area under Precision-Recall Curve (AUPRC) 0.9848 test set. Among CG-JKNN top attaining 0.9549 AUPRC 0.9846 held-out integration morphological features proved highly effective, with XGBoost showing promising results classifying instances AFB nucleus. identified closely align criteria used pathologists practice, highlighting clinical applicability approach. Future work will explore distillation techniques graph-level classification distinct progression categories.

Язык: Английский

---

Analyzing the performance of a bio-sensor integrated improved blended learning model for accurate pneumonia prediction

Results in Engineering, Год журнала: 2024, Номер 22, С. 102063 - 102063

Опубликована: Апрель 2, 2024

Pneumonia has been considered a life-threatening disease for elderly human beings and those with weakened immune systems in the present medical era. The contemporary scenario highlights significance of intelligent automatic handheld devices to detect pneumonia other pulmonary diseases. Hence, this research designed an improved blended learning paradigm (IBLP) real-time detection from chest X-rays, early lung diseases alveolar gas using biosensors graphical processing unit (GPU) developed overcome resolve such challenges. It emphasizes applications techniques, particularly identifying X-ray images exhaled breath support vector machine (SVM). experimental findings indicate that based VGG16 (91.99%) consistently outperforms VGG19 (88.91%) ResNet50 (87.02%) model diagnostic accuracy. IBLP provided 95.5% precision, 97.69% F1 score, 100% recall rate no false-negative results. future classification diagnosis will likely involve artificial intelligence-based can provide accurate timely analysis images, thereby improving patient outcomes reducing healthcare costs.

Язык: Английский

---

Optimizing lung cancer prediction: leveraging Kernel PCA with dendritic neural models

Computer Methods in Biomechanics & Biomedical Engineering, Год журнала: 2024, Номер unknown, С. 1 - 14

Опубликована: Июль 13, 2024

Lung cancer is considered a cause of increased mortality rate due to delays in diagnostics. There an urgent need develop effective lung prediction model that will help the early diagnosis and save patients from unnecessary treatments. The objective current paper meet extensiveness measure by using collaborative feature selection extraction methods enhance dendritic neural (DNM) comparison traditional machine learning (ML) models with minimum features boost accuracy, precision, sensitivity prediction. Comprehensive experiments on dataset comprising 1000 23 obtained Kaggle. Crucial are identified, proposed method's effectiveness evaluated metrics such as F1 score, sensitivity, specificity, confusion matrix against other ML models. Feature techniques including Principal Component Analysis (PCA), Kernel PCA (K-PCA), Uniform Manifold Approximation Projection (UMAP) employed optimize performance. DNM accuracy at 96.50%, precision 96.64% 97.45% sensitivity. K-PCA explained 98.50%, 99.42%, 98.84% UMAP elaborated 98%, 98.82%, 98.82% approach showed outstanding performance enhancing model. Highlighting DNM's accurate cancer. These results emphasize potential contribute positively healthcare research providing better predictive outcomes.

Язык: Английский

---

Edge-Cloud Synergy for AI-Enhanced Sensor Network Data: A Real-Time Predictive Maintenance Framework

Sensors, Год журнала: 2024, Номер 24(24), С. 7918 - 7918

Опубликована: Дек. 11, 2024

Sensor networks generate vast amounts of data in real-time, which challenges existing predictive maintenance frameworks due to high latency, energy consumption, and bandwidth requirements. This research addresses these limitations by proposing an edge-cloud hybrid framework, leveraging edge devices for immediate anomaly detection cloud servers in-depth failure prediction. A K-Nearest Neighbors (KNNs) model is deployed on detect anomalies reducing the need continuous transfer cloud. Meanwhile, a Long Short-Term Memory (LSTM) analyzes time-series analysis, enhancing scheduling operational efficiency. The framework’s dynamic workload management algorithm optimizes task distribution between resources, balancing usage, consumption. Experimental results show that approach achieves 35% reduction 28% decrease 60% usage compared cloud-only solutions. framework offers scalable, efficient solution real-time maintenance, making it highly applicable resource-constrained, data-intensive environments.

Язык: Английский

---