Deep learning models in classifying primary bone tumors and bone infections based on radiographs

npj Precision Oncology, Год журнала: 2025, Номер 9(1)

Опубликована: Март 13, 2025

Primary bone tumors (PBTs) present significant diagnostic challenges due to their heterogeneous nature and similarities with infections. This study aimed develop an ensemble deep learning framework that integrates multicenter radiographs extensive clinical features accurately differentiate between PBTs We compared the performance of model four imaging models based solely on utilizing EfficientNet B3, B4, Vision Transformer, Swin Transformers. The patients were split into external dataset (N = 423) internal [including training 1044), test 354), validation set 171)]. outperformed models, achieving areas under curve (AUCs) 0.948 0.963 sets, respectively, accuracies 0.881 0.895. Its surpassed junior mid-level radiologists was comparable senior (accuracy: 83.6%). These findings underscore potential in enhancing precision for infections (Research Registration Unique Identifying Number (UIN): researchregistry10483 details are available at https://www.researchregistry.com/register-now#home/registrationdetails/6693845995ba110026aeb754/ ).

Язык: Английский

---

Deep learning-assisted medical image compression challenges and opportunities: systematic review

Neural Computing and Applications, Год журнала: 2024, Номер 36(17), С. 10067 - 10108

Опубликована: Апрель 30, 2024

Язык: Английский

---

An Adaptation of Hybrid Binary Optimization Algorithms for Medical Image Feature Selection in Neural Network for Classification of Breast Cancer

Neurocomputing, Год журнала: 2024, Номер unknown, С. 129018 - 129018

Опубликована: Ноя. 1, 2024

Язык: Английский

---

EnhanceDeepIris Model for Iris Recognition Applications

IEEE Access, Год журнала: 2024, Номер 12, С. 66809 - 66821

Опубликована: Янв. 1, 2024

In this study, an iris recognition technique based on enhanced EnhanceDeepIris model is proposed in order to examine at a deeper level. The process first uses convolutional features extract human features. Then, solve the deformation problem caused by radial data of texture, sequential metric introduced realize effective iris. Finally, structure thoroughly examined investigate its actual performance. study found that after running four algorithms ND-IRIS-0405 and CASIA-Lamp datasets, loss function value research method began approach lowest 14th 9th iterations, while other continued slowly decrease. Additionally, 5th 4th iterations respectively, accuracy was nearly 91.00%.After 20 classification predictions, average accuracies paper, segmentation end-to-end multi-task network IrisST-Net, full complex-valued neural network, with hybrid preprocessing feature extraction were 99.88%, 98.72%, 97.47%, 89.77%, respectively. These findings suggest recognizes best can identify primary contour information accuracy. results application demonstrate provides clearest detection edge, less noise, accurately detect main This reference for optimizing related technologies field image recognition.

Язык: Английский

---

AI in infectious disease diagnosis and vaccine development

Methods in microbiology, Год журнала: 2025, Номер unknown

Опубликована: Янв. 1, 2025

Язык: Английский

---

MCU-Net: A multi-prior collaborative deep unfolding network with gates-controlled spatial attention for accelerated MRI reconstruction

Neurocomputing, Год журнала: 2025, Номер 633, С. 129771 - 129771

Опубликована: Фев. 26, 2025

Язык: Английский

---

A Novel Deep Learning model for detection of Pneumonia and Covid-19 variants from Chest X-ray images

International Journal of Computational and Experimental Science and Engineering, Год журнала: 2025, Номер 11(2)

Опубликована: Апрель 13, 2025

Pneumonia is an example of a past pandemic and continues to be serious health concern. In the USA, more than one million people are admitted in hospital with pneumonia every year, leading about 500,000 deaths. Chest X-ray imaging effective widely utilised method for diagnosing essential both healthcare epidemiological studies. COVID-19, viral infection initiated Wuhan, China towards end 2019, quickly spread across globe. It caused by Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) has influenced millions globally. Analyzing images regarded as fastest simplest methods discovery, available at minimal cost many places. CT scans, on other way, mere advanced technique that can identify small changes composition internal organs. This uses 3-D computer technology along X-rays detailed examination. While scans provide body compositions, traditional sometimes occlude, making it difficult see fine details. The proposed model outlines framework classifying COVID-19 variants predicting new ones. As per results, ResNet_Seg achieved F1 score 99.96%, which higher CNN models tested. performance these assessed using datasets from SARS MERS, resulting accurate predictions. Future work will focus validating statistical methods. A relative analysis deep learning models, including CNN, ResNet, Darknet, conducted, enhancements through novel segmentation algorithm hyperparameter fine-tuning. results offer insights into developing reliable diagnostic methodologies machine techniques.

Язык: Английский

---

Boundary guided network with two-stage transfer learning for gastrointestinal polyps segmentation

Expert Systems with Applications, Год журнала: 2023, Номер 240, С. 122503 - 122503

Опубликована: Ноя. 8, 2023

Язык: Английский

---

Concatenated CNN-Based Pneumonia Detection Using a Fuzzy-Enhanced Dataset

Sensors, Год журнала: 2024, Номер 24(20), С. 6750 - 6750

Опубликована: Окт. 21, 2024

Pneumonia is a form of acute respiratory infection affecting the lungs. Symptoms viral and bacterial pneumonia are similar. Rapid diagnosis disease difficult, since polymerase chain reaction-based methods, which have greatest reliability, provide results in few hours, while ensuring high requirements for compliance with analysis technology professionalism personnel. This study proposed Concatenated CNN model detection combined fuzzy logic-based image improvement method. The enhancement process based on new fuzzification refinement algorithm, significantly improved quality feature extraction CCNN model. Four datasets, original upgraded images utilizing entropy, standard deviation, histogram equalization, were utilized to train algorithm. CCNN's performance was demonstrated be by entropy-added dataset producing best results. suggested attained remarkable classification metrics, including 98.9% accuracy, 99.3% precision, 99.8% F1-score, 99.6% recall. Experimental comparisons showed that worked better than traditional resulting higher diagnostic precision. demonstrates how well deep learning models sophisticated techniques work together analyze medical images.

Язык: Английский

---

Classifying chest x-rays for COVID-19 through transfer learning: a systematic review

Multimedia Tools and Applications, Год журнала: 2024, Номер unknown

Опубликована: Апрель 2, 2024

Язык: Английский

---