Scholarlink

<p>Breast cancer remains one of the most prevalent and life-threatening diseases affecting women worldwide. Early detection is critical for improving prognosis and survival rates, making it a key focus in current research. This review systematically examines over 100 recent studies related to breast cancer detection and diagnosis, with an emphasis on machine and deep learning applications. We propose a new classification approach based on the type of data used—such as histopathological images, mammograms, ultrasound, and multi-modal data—rather than traditional method-based categorizations. For each data type, we analyze employed techniques, preprocessing steps, feature selection strategies, classification models, and evaluation metrics. Relevant screening filters and performance measures are also highlighted to present a holistic view of the state-of-the-art. The findings reveal that data-driven classification not only unifies diverse techniques but also highlights emerging trends and research gaps in each data category. This review aims to serve as a comprehensive and practical guide for researchers and clinicians working in breast cancer imaging and AI-assisted diagnosis.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The role of pre-processing, machine, and deep learning approaches in advancing breast cancer diagnosis and detection using different data modalities: a review

  • 期刊论文

  • 发表日期：

    2025年8月4日

查看全文
  • Saida Sarra Boudouh,
  • Mustapha Bouakkaz

摘要

Breast cancer remains one of the most prevalent and life-threatening diseases affecting women worldwide. Early detection is critical for improving prognosis and survival rates, making it a key focus in current research. This review systematically examines over 100 recent studies related to breast cancer detection and diagnosis, with an emphasis on machine and deep learning applications. We propose a new classification approach based on the type of data used—such as histopathological images, mammograms, ultrasound, and multi-modal data—rather than traditional method-based categorizations. For each data type, we analyze employed techniques, preprocessing steps, feature selection strategies, classification models, and evaluation metrics. Relevant screening filters and performance measures are also highlighted to present a holistic view of the state-of-the-art. The findings reveal that data-driven classification not only unifies diverse techniques but also highlights emerging trends and research gaps in each data category. This review aims to serve as a comprehensive and practical guide for researchers and clinicians working in breast cancer imaging and AI-assisted diagnosis.