Enhancing pediatric pneumonia diagnosis through masked autoencoders

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: March 14, 2024

Abstract Pneumonia, an inflammatory lung condition primarily triggered by bacteria, viruses, or fungi, presents distinctive challenges in pediatric cases due to the unique characteristics of respiratory system and potential for rapid deterioration. Timely diagnosis is crucial, particularly children under 5, who have immature immune systems, making them more susceptible pneumonia. While chest X-rays are indispensable diagnosis, arise from subtle radiographic findings, varied clinical presentations, subjectivity interpretations, especially cases. Deep learning, transfer has shown promise improving pneumonia leveraging large labeled datasets. However, scarcity data a hurdle effective model training. To address this challenge, we explore self-supervised focusing on Masked Autoencoder (MAE). By pretraining MAE adult X-ray images fine-tuning pretrained dataset, aim overcome issues enhance diagnostic accuracy The proposed approach demonstrated competitive performance AUC 0.996 95.89% distinguishing between normal Additionally, exhibited high values (normal: 0.997, bacterial pneumonia: 0.983, viral 0.956) 93.86% classifying normal, pneumonia, This study also investigated impact different masking ratios during explored efficiency model, presenting enhanced capabilities

Language: Английский

---

A deep ensemble learning framework for COVID-19 detection in chest X-ray images

Network Modeling Analysis in Health Informatics and Bioinformatics, Journal Year: 2024, Volume and Issue: 13(1)

Published: May 31, 2024

Language: Английский

---

Adaptive loss optimization for enhanced learning performance: application to image-based rock classification

Neural Computing and Applications, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 9, 2025

Language: Английский

---

PediaPulmoDx: Harnessing Cutting Edge Preprocessing and Explainable AI for Pediatric Chest X-ray Classification with DenseNet121

Results in Engineering, Journal Year: 2025, Volume and Issue: unknown, P. 104320 - 104320

Published: Feb. 1, 2025

Language: Английский

---

The Diagnostic Classification of the Pathological Image Using Computer Vision

Algorithms, Journal Year: 2025, Volume and Issue: 18(2), P. 96 - 96

Published: Feb. 8, 2025

Computer vision and artificial intelligence have revolutionized the field of pathological image analysis, enabling faster more accurate diagnostic classification. Deep learning architectures like convolutional neural networks (CNNs), shown superior performance in tasks such as classification, segmentation, object detection pathology. has significantly improved accuracy disease diagnosis healthcare. By leveraging advanced algorithms machine techniques, computer systems can analyze medical images with high precision, often matching or even surpassing human expert performance. In pathology, deep models been trained on large datasets annotated pathology to perform cancer diagnosis, grading, prognostication. While approaches show great promise challenges remain, including issues related model interpretability, reliability, generalization across diverse patient populations imaging settings.

Language: Английский

---

A Segmentation-Based Approach for Lung Disease Classification Using Chest X-ray Images

Communications in computer and information science, Journal Year: 2025, Volume and Issue: unknown, P. 172 - 185

Published: Jan. 1, 2025

Language: Английский

---

LungNet-ViT: Efficient lung disease classification using a multistage vision transformer model from chest radiographs

Journal of X-Ray Science and Technology, Journal Year: 2025, Volume and Issue: unknown

Published: March 28, 2025

This research introduces a Multistage-Vision Transformer (Multistage-ViT) model for precisely classifying various lung diseases using chest radiographic (CXR) images. The dataset in the proposed method includes four classes: Normal, COVID-19, Viral Pneumonia and Lung Opacity. demonstrates its efficacy on imbalanced balanced datasets by enhancing classifier accuracy through deep feature extraction. It integrates backbone models with ViT architecture, creating rigorously hybrid configurations compared to their standalone counterparts. These utilize optimized features classification, significantly improving performance. Notably, multistage-ViT achieved accuracies of 99.93% an 99.97% InceptionV3 combined model. findings highlight superior robustness models, underscoring potential enhance disease classification advanced extraction integration techniques. effectively benefits employing from CXR

Language: Английский

---

Optimizing Pneumonia Identification in Chest X-Rays Using Deep Learning Pre-Trained Architecture for Image Reconstruction in Medical Imaging

International Journal of Advanced Research in Science Communication and Technology, Journal Year: 2025, Volume and Issue: unknown, P. 52 - 63

Published: April 2, 2025

The rapid accumulation of fluid in the lungs is hallmark fatal illness known as pneumonia. Therefore, it crucial to get a diagnosis and medication soon possible order stop condition from getting worse. In diagnose pneumonia, chest X-rays (CXR) are typically used. This study assesses efficacy ResNet50 other pre-trained DL models classifying X-ray images evidence proposed model achieves an accuracy 93.06%, precision 88.97%, recall 96.78%, F1-score 92.71%, surpassing MobileNet, EfficientNetB0, Xception across all performance metrics. has been further tested shown be reliable effective differentiating between normal pneumonia patients utilizing ROC curve analysis, accuracy-loss trends, confusion matrix. highlights superiority automated detection, offering promising tool for early clinical decision support. results highlight how deep learning-based methods have ability improve radiological evaluations, which turn can decrease diagnostic mistakes increase patient outcomes. research contributes developing AI-driven medical imaging solutions, facilitating more accurate scalable detection real-world healthcare settings

Language: Английский

---

An effective PO-RSNN and FZCIS based diabetes prediction and stroke analysis in the metaverse environment

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: April 4, 2025

Abstract Chronic disease (CD) like diabetes and stroke impacts global healthcare extensively, continuous monitoring early detection are necessary for effective management. The Metaverse Environment (ME) has gained attention in the digital environment; yet, it lacks adequate support disabled individuals, including deaf dumb people, also faces challenges security, generalizability, feature selection. To overcome these limitations, a novel probabilistic-centric optimized recurrent sechelliott neural network (PO-RSNN)-based prediction (DP) Fuzzy Z-log-clipping inference system (FZCIS)-based severity level estimation ME is carried out. proposed integrates Montwisted-Jaco curve cryptography (MJCC) secured data transmission, Aransign-principal component analysis (A-PCA) dimensionality reduction, synthetic minority oversampling technique (SMOTE) to address imbalance. diagnosed results securely stored BlockChain (BC) enhanced privacy traceability. experimental validation demonstrated superior performance of by achieving 98.97% accuracy DP 98.89% analysis, outperforming existing classifiers. Also, MJCC attained 98.92% efficiency, surpassing traditional encryption models. Thus, produces secure, scalable, highly accurate ME. Further, research will extend approach other CD cancer heart improve predictive performance.

Language: Английский

---

A Distillation Approach to Transformer-Based Medical Image Classification with Limited Data

Diagnostics, Journal Year: 2025, Volume and Issue: 15(7), P. 929 - 929

Published: April 4, 2025

Background/Objectives: Although transformer-based deep learning architectures are preferred in many hybrid due to their flexibility, they generally perform poorly on image classification tasks with small datasets. An important improvement performance when transformer work limited data is the use of distillation techniques. The impact techniques accuracy models has not yet been extensively investigated. Methods: This study investigates trained data. We ViTx32 and ViTx16 without DeiT BeiT distillation. A four-class dataset brain MRI images used for training testing. Results: Our experiments show that achieve gains 2.2% 1%, respectively, compared ViTx16. more detailed analysis shows improve detection non-patient individuals by about 4%. also includes a times each architecture. Conclusions: results using can significantly working Based these findings, we recommend distillation, especially medical applications other areas where flexible developed

Language: Английский

---