Enhancing Breast Cancer Detection in Ultrasound Images: An Innovative Approach Using Progressive Fine‐Tuning of Vision Transformer Models DOI Creative Commons
Meshrif Alruily,

Alshimaa Abdelraof Mahmoud,

Hisham Allahem

et al.

International Journal of Intelligent Systems, Journal Year: 2024, Volume and Issue: 2024(1)

Published: Jan. 1, 2024

Breast cancer is ranked as the second most common among women globally, highlighting critical need for precise and early detection methods. Our research introduces a novel approach classifying benign malignant breast ultrasound images. We leverage advanced deep learning methodologies, mainly focusing on vision transformer (ViT) model. method distinctively features progressive fine‐tuning, tailored process that incrementally adapts model to nuances of tissue classification. Ultrasound imaging was chosen its distinct benefits in medical diagnostics. This modality noninvasive cost‐effective demonstrates enhanced specificity, especially dense tissues where traditional methods may struggle. Such characteristics make it an ideal choice sensitive task detection. extensive experiments utilized images dataset, comprising 780 both tissues. The dataset underwent comprehensive analysis using several pretrained models, including VGG16, VGG19, DenseNet121, Inception, ResNet152V2, DenseNet169, DenseNet201, ViT. results presented were achieved without employing data augmentation techniques. ViT demonstrated robust accuracy generalization capabilities with original size, which consisted 637 Each model’s performance meticulously evaluated through 10‐fold cross‐validation technique, ensuring thorough unbiased comparison. findings are significant, demonstrating fine‐tuning substantially enhances capability. resulted remarkable 94.49% AUC score 0.921, significantly higher than models fine‐tuning. These affirm efficacy highlight transformative potential integrating image classification tasks. study solidifies role such methodologies improving diagnosis, when coupled unique advantages imaging.

Language: Английский

Fluorescence microscopy and histopathology image based cancer classification using graph convolutional network with channel splitting DOI
Asish Bera, Debotosh Bhattacharjee, Ondřej Krejcar

et al.

Biomedical Signal Processing and Control, Journal Year: 2025, Volume and Issue: 103, P. 107400 - 107400

Published: Jan. 6, 2025

Language: Английский

Citations

0
A Deep Learning with Metaheuristic Optimization-Driven Breast Cancer Segmentation and Classification Model using Mammogram Imaging DOI Open Access

M. Sreevani,

R. Latha

Engineering Technology & Applied Science Research, Journal Year: 2025, Volume and Issue: 15(1), P. 20342 - 20347

Published: Feb. 2, 2025

Cancer is the second leading cause of death globally, with Breast (BC) accounting for 20% new diagnoses, making it a major morbidity and mortality. Mammography effective BC detection, but lesion interpretation challenging, prompting development Computer-Aided Diagnosis (CAD) systems to assist in classification detection. Machine Learning (ML) Deep (DL) models are widely used disease diagnosis. Therefore, this study presents an Optimized Graph Convolutional Recurrent Neural Network based Segmentation Recognition Classification (OGCRNN-SBCRC) technique. In preparation phase, images masks annotated then classified as benign or malignant. To achieve this, Wiener Filter (WF)-based noise removal log transform-based contrast enhancement preprocessing. The OGCRNN-SBCRC technique utilizes UNet++ method segmentation RMSProp optimizer parameter tuning. addition, employs ConvNeXtTiny Convolution (CNN) approach feature extraction. For (GCRNN) model used. Finally, Aquila Optimizer (AO) employed hyperparameter tuning GCRNN approach. simulation analysis methodology, using image dataset, demonstrated superior performance accuracy 99.65%, surpassing existing models.

Language: Английский

Citations

0
Advanced Hybridization and Optimization of DNNs for Medical Imaging: A Survey on Disease Detection Techniques DOI Creative Commons

Maneet Kaur Bohmrah,

Harjot Kaur

Artificial Intelligence Review, Journal Year: 2025, Volume and Issue: 58(4)

Published: Feb. 4, 2025

Language: Английский

Citations

0
Advanced computational techniques: Bridging metaheuristic optimization and deep learning for material design through image enhancement DOI
Jagrati Talreja,

Divya Chauhan

Elsevier eBooks, Journal Year: 2025, Volume and Issue: unknown, P. 197 - 228

Published: Jan. 1, 2025

Language: Английский

Citations

0
Multi-condition pipeline leak diagnosis based on acoustic image fusion and whale-optimized evolutionary convolutional neural network DOI
Yuan Yuan, Xiwang Cui, Xiaojuan Han

et al.

Engineering Applications of Artificial Intelligence, Journal Year: 2025, Volume and Issue: 153, P. 110886 - 110886

Published: April 22, 2025

Language: Английский

Citations

0
A hybrid deep learning model for mammographic breast cancer detection: Multi-autoencoder and attention mechanisms DOI
Long Yan, Lei Wu, Meng Xia

et al.

Journal of Radiation Research and Applied Sciences, Journal Year: 2025, Volume and Issue: 18(3), P. 101578 - 101578

Published: May 8, 2025

Language: Английский

Citations

0
Advanced Breast Cancer Diagnostics With PolyBreastVit: A Combined PolyNet and Vision Transformer Approach DOI Creative Commons
Visalakshi Annepu, Mohamed Abbas, Hanumantharao Bitra

et al.

Applied Computational Intelligence and Soft Computing, Journal Year: 2024, Volume and Issue: 2024(1)

Published: Jan. 1, 2024

Breast cancer continues to be an important health issue around the world, with timely screening being in improving survival and therapy. Here is a presentation of PolyBreastVit, novel hybrid deep learning (DL) model for automatic detection classification breast ultrasound images that combines PolyNet Vision Transformer (ViT). The above trained validated on dataset 880 high‐definition collected from 500 female subjects aged between 25 75 years three classes: benign, malignant, normal. For enhancement proposed model’s accuracy, thorough data augmentation preprocessing have been performed. performance PolyBreastVit evaluated against several well‐known DL models such as VGG‐16, Inception V3, ResNet‐50 using precision, recall, F 1, AUC, other standard metrics. These findings support evidence manages outperform those classical task every aspect. This paper presents latest development diagnostic tools through medical imaging incorporating convolutional neural networks (CNNs) transformer radiologists.

Language: Английский

Citations

0
Enhancing Breast Cancer Detection in Ultrasound Images: An Innovative Approach Using Progressive Fine‐Tuning of Vision Transformer Models DOI Creative Commons
Meshrif Alruily,

Alshimaa Abdelraof Mahmoud,

Hisham Allahem

et al.

International Journal of Intelligent Systems, Journal Year: 2024, Volume and Issue: 2024(1)

Published: Jan. 1, 2024

Breast cancer is ranked as the second most common among women globally, highlighting critical need for precise and early detection methods. Our research introduces a novel approach classifying benign malignant breast ultrasound images. We leverage advanced deep learning methodologies, mainly focusing on vision transformer (ViT) model. method distinctively features progressive fine‐tuning, tailored process that incrementally adapts model to nuances of tissue classification. Ultrasound imaging was chosen its distinct benefits in medical diagnostics. This modality noninvasive cost‐effective demonstrates enhanced specificity, especially dense tissues where traditional methods may struggle. Such characteristics make it an ideal choice sensitive task detection. extensive experiments utilized images dataset, comprising 780 both tissues. The dataset underwent comprehensive analysis using several pretrained models, including VGG16, VGG19, DenseNet121, Inception, ResNet152V2, DenseNet169, DenseNet201, ViT. results presented were achieved without employing data augmentation techniques. ViT demonstrated robust accuracy generalization capabilities with original size, which consisted 637 Each model’s performance meticulously evaluated through 10‐fold cross‐validation technique, ensuring thorough unbiased comparison. findings are significant, demonstrating fine‐tuning substantially enhances capability. resulted remarkable 94.49% AUC score 0.921, significantly higher than models fine‐tuning. These affirm efficacy highlight transformative potential integrating image classification tasks. study solidifies role such methodologies improving diagnosis, when coupled unique advantages imaging.

Language: Английский

Citations

0