RADIC:A tool for diagnosing COVID-19 from chest CT and X-ray scans using deep learning and quad-radiomics

Chemometrics and Intelligent Laboratory Systems, Год журнала: 2023, Номер 233, С. 104750 - 104750

Опубликована: Янв. 2, 2023

Язык: Английский

---

Glaucoma diagnosis from fundus images using modified Gauss-Kuzmin-distribution-based Gabor features in 2D-FAWT

Computers & Electrical Engineering, Год журнала: 2024, Номер 119, С. 109538 - 109538

Опубликована: Авг. 22, 2024

Язык: Английский

---

Enhanced Multi-Model Deep Learning for Rapid and Precise Diagnosis of Pulmonary Diseases Using Chest X-Ray Imaging

Diagnostics, Год журнала: 2025, Номер 15(3), С. 248 - 248

Опубликована: Янв. 22, 2025

Background: The global burden of respiratory diseases such as influenza, tuberculosis, and viral pneumonia necessitates rapid, accurate diagnostic tools to improve healthcare responses. Current methods, including RT-PCR chest radiography, face limitations in accuracy, speed, accessibility, cost-effectiveness, especially resource-constrained settings, often delaying treatment increasing transmission. Methods: This study introduces an Enhanced Multi-Model Deep Learning (EMDL) approach address these challenges. EMDL integrates ensemble five pre-trained deep learning models (VGG-16, VGG-19, ResNet, AlexNet, GoogleNet) with advanced image preprocessing (histogram equalization contrast enhancement) a novel multi-stage feature selection optimization pipeline (PCA, SelectKBest, Binary Particle Swarm Optimization (BPSO), Grey Wolf (BGWO)). Results: Evaluated on two independent X-ray datasets, achieved high accuracy the multiclass classification pneumonia, tuberculosis. combined enhancement strategies significantly improved precision model robustness. Conclusions: framework provides scalable efficient solution for accessible pulmonary disease diagnosis, potentially improving efficacy patient outcomes, particularly resource-limited settings.

Язык: Английский

---

Comparison of CNN Classification Model using Machine Learning with Bayesian Optimizer

HighTech and Innovation Journal, Год журнала: 2023, Номер 4(3), С. 531 - 542

Опубликована: Сен. 1, 2023

One of the best-known and frequently used areas Deep Learning in image processing is Convolutional Neural Network (CNN), which has architectural designs such as Inceptionv3, DenseNet201, Resnet50, MobileNet classification pattern recognition. Furthermore, CNN extracts feature from according to designed architecture performs through fully connected layer, executes Machine (ML) algorithm tasks. Examples ML that are commonly include Naive Bayes (NB), k-Nearest Neighbor (k-NN), Support Vector (SVM), Decision Tree (DT). This research was conducted based on an AI model development background need for a system diagnose COVID-19 quickly accurately. The aim classify aforementioned models with algorithms compare models’ accuracy before after Bayesian optimization using CXR lung images total 2000 data. Consequently, extracted 80% training data 20% testing, assigned four different use ensure best accuracy. It observed generated by MobileNetV2-SVM structure 93%. Therefore, obtained SVM higher than other three algorithms. Doi: 10.28991/HIJ-2023-04-03-05 Full Text: PDF

Язык: Английский

---

Hair and Scalp Disease Detection using Machine Learning and Image Processing

European Journal of Information Technologies and Computer Science, Год журнала: 2023, Номер 3(1), С. 7 - 13

Опубликована: Янв. 23, 2023

Almost 80 million Americans suffer from hair loss due to aging, stress, medication, or genetic makeup. Hair and scalp-related diseases often go unnoticed in the beginning. Sometimes, a patient cannot differentiate between regular fall. Diagnosing hair-related is time-consuming as it requires professional dermatologists perform visual medical tests. Because of that, overall diagnosis gets delayed, which worsens severity illness. Due image-processing ability, neural network-based applications are used various sectors, especially healthcare health informatics, predict deadly like cancers tumors. These assist clinicians patients provide an initial insight into early-stage symptoms. In this study, we deep learning approach that successfully predicts three main types diseases: alopecia, psoriasis, folliculitis. However, limited study area, unavailability proper dataset, degree variety among images scattered over internet made task challenging. 150 were obtained sources then preprocessed by denoising, image equalization, enhancement, data balancing, thereby minimizing error rate. After feeding processed 2D convolutional network (CNN) model, training accuracy 96.2%, with validation 91.1%. The precision recall score folliculitis 0.895, 0.846, 1.0, respectively. We also created dataset scalp for future prospective researchers.

Язык: Английский

---

An Efficient and Robust Approach Using Inductive Transfer-Based Ensemble Deep Neural Networks for Kidney Stone Detection

IEEE Access, Год журнала: 2024, Номер 12, С. 32894 - 32910

Опубликована: Янв. 1, 2024

Chronic kidney disorder is a global health problem involving the repercussions of impaired function and failure. A stone scenario that impairs function. Because this disease usually asymptomatic, early quick detection problems essential to avoid significant consequences. This study presents an automated Computed Tomography (CT) images using inductive transfer-based ensemble Deep Neural Network (DNN). Three datasets are created for feature extraction from CT pre-trained DNN models. After assembling several DNNs, such as DarkNet19, InceptionV3, ResNet101, deep vector concatenation. The Iterative ReliefF selection method used choose most informative vectors, which then fed into K Nearest Neighbor classifier tuned Bayesian optimizer with 10-fold cross-validation approach detect stones. proposed strategy achieves 99.8% 96.7% accuracy quality noisy image datasets, superior other DNN-based traditional approaches. can help urologists confirm their physical inspection stones, reducing possibility human mistakes.

Язык: Английский

---

A hybridized feature extraction for COVID-19 multi-class classification on computed tomography images

Heliyon, Год журнала: 2024, Номер 10(5), С. e26939 - e26939

Опубликована: Фев. 24, 2024

COVID-19 has killed more than 5 million individuals worldwide within a short time. It is caused by SARS-CoV-2 which continuously mutates and produces transmissible new different strains. therefore of great significance to diagnose early curb its spread reduce the death rate. Owing pandemic, traditional diagnostic methods such as reverse-transcription polymerase chain reaction (RT-PCR) are ineffective for diagnosis. Medical imaging among most effective techniques respiratory disorders detection through machine learning deep learning. However, conventional depend on extracted engineered features, whereby optimum features influence classifier's performance. In this study, Histogram Oriented Gradient (HOG) eight models were utilized feature extraction while K-Nearest Neighbour (KNN) Support Vector Machines (SVM) used classification. A combined HOG was proposed improve performance classifiers. VGG-16 + achieved 99.4 overall accuracy with SVM. This indicates that our concatenated can enhance SVM in detection.

Язык: Английский

---

Combating the COVID-19 infodemic using Prompt-Based curriculum learning

Expert Systems with Applications, Год журнала: 2023, Номер 229, С. 120501 - 120501

Опубликована: Май 18, 2023

Язык: Английский

---

Analysis for diagnosis of pneumonia symptoms using chest X-ray based on MobileNetV2 models with image enhancement using white balance and contrast limited adaptive histogram equalization (CLAHE)

Biomedical Signal Processing and Control, Год журнала: 2023, Номер 90, С. 105857 - 105857

Опубликована: Дек. 15, 2023

Язык: Английский

---

Swin-textural: A novel textural features-based image classification model for COVID-19 detection on chest computed tomography

Informatics in Medicine Unlocked, Год журнала: 2022, Номер 36, С. 101158 - 101158

Опубликована: Дек. 31, 2022

Chest computed tomography (CT) has a high sensitivity for detecting COVID-19 lung involvement and is widely used diagnosis disease monitoring. We proposed new image classification model, swin-textural, that combined swin-based patch division with textual feature extraction automated of on chest CT images. The main objective this work to evaluate the performance swin architecture in engineering.

Язык: Английский

---