Early Breast Cancer Detection Based on Deep Learning: An Ensemble Approach Applied to Mammograms DOI Creative Commons
Youness Khourdifi, Alae El Alami,

Mounia Zaydi

и другие.

BioMedInformatics, Год журнала: 2024, Номер 4(4), С. 2338 - 2373

Опубликована: Дек. 13, 2024

Background: Breast cancer is one of the leading causes death in women, making early detection through mammography crucial for improving survival rates. However, human interpretation mammograms often prone to diagnostic errors. This study addresses challenge accuracy breast by leveraging advanced machine learning techniques. Methods: We propose an extended ensemble deep model that integrates three state-of-the-art convolutional neural network (CNN) architectures: VGG16, DenseNet121, and InceptionV3. The utilizes multi-scale feature extraction enhance both benign malignant masses mammograms. approach evaluated on two benchmark datasets: INbreast CBIS-DDSM. Results: proposed achieved significant performance improvements. On dataset, attained 90.1%, recall 88.3%, F1-score 89.1%. For CBIS-DDSM reached 89.5% 90.2% specificity. method outperformed each individual CNN model, reducing false positives negatives, thereby providing more reliable results. Conclusions: demonstrated strong potential as a decision support tool radiologists, offering accurate earlier cancer. By complementary strengths multiple architectures, this can improve clinical accessibility high-quality screening.

Язык: Английский

Unravelling Data Challenges in AI-Driven Alzheimer's Research DOI

B. Arunadevi,

Vidyabharathi Dakshinamurthi,

Bennilo Fernandes

и другие.

Advances in medical technologies and clinical practice book series, Год журнала: 2024, Номер unknown, С. 211 - 222

Опубликована: Июнь 28, 2024

Alzheimer's disease (AD) is a rapidly developing public fitness subject, affecting thousands and of human beings globally placing sizable strain on healthcare systems. With the upward push synthetic intelligence (AI) technologies, there was renewed interest in using records-driven approaches to apprehend potentially deal with advert. In this chapter, authors aim get bottom these data challenges AI-pushed studies, exploring ability solutions destiny instructions. They first speak about various forms used AD studies. then examine common facts best troubles biases that can have an effect AI fashions, recommend processes mitigate those demanding situations. end, they capability collaborative statistics-sharing projects conquer advance AI-driven Through information addressing challenges, pave way for greater correct impactful fight against devastating disease.

Язык: Английский

Процитировано

0
Ethics & Responsible AI in Healthcare DOI Creative Commons

Fardin Quazi

Опубликована: Авг. 13, 2024

Artificial Intelligence, Healthcare, Ethics, Responsible AI, Diagnostic Treatment Planning, Patient Care, Governance Frameworks, Machine Learning, Data Privacy, Safety, Predictive Analysis, Decision Support Systems, Future of AI in Healthcare.

Язык: Английский

Процитировано

0
Early Breast Cancer Detection Based on Deep Learning: An Ensemble Approach Applied to Mammograms DOI Creative Commons
Youness Khourdifi, Alae El Alami,

Mounia Zaydi

и другие.

BioMedInformatics, Год журнала: 2024, Номер 4(4), С. 2338 - 2373

Опубликована: Дек. 13, 2024

Background: Breast cancer is one of the leading causes death in women, making early detection through mammography crucial for improving survival rates. However, human interpretation mammograms often prone to diagnostic errors. This study addresses challenge accuracy breast by leveraging advanced machine learning techniques. Methods: We propose an extended ensemble deep model that integrates three state-of-the-art convolutional neural network (CNN) architectures: VGG16, DenseNet121, and InceptionV3. The utilizes multi-scale feature extraction enhance both benign malignant masses mammograms. approach evaluated on two benchmark datasets: INbreast CBIS-DDSM. Results: proposed achieved significant performance improvements. On dataset, attained 90.1%, recall 88.3%, F1-score 89.1%. For CBIS-DDSM reached 89.5% 90.2% specificity. method outperformed each individual CNN model, reducing false positives negatives, thereby providing more reliable results. Conclusions: demonstrated strong potential as a decision support tool radiologists, offering accurate earlier cancer. By complementary strengths multiple architectures, this can improve clinical accessibility high-quality screening.

Язык: Английский

Процитировано

0