Scholarlink

Breast cancer (BC) is a condition where aberrant breast cells proliferate uncontrollably, resulting in tumors, and the most hazardous disease for women worldwide is BC. It frequently results in severe psychological suffering, including elevated rates of anxiety and depression. Machine learning (ML) and computer-aided diagnosis (CAD) have demonstrated potential in enhancing the identification of BC, specifically in the area of Deep Learning (DL) methods for the classification of histopathological images (HI). These techniques depend on huge amounts of labeled data, which can be challenging to gather. This study presents a new method called Aquila Optimizer (AO) with Hybrid ResNet-DenseNet Enabled Breast Cancer Classification (HRD-BC2HI), combining ResNet and DenseNet models for feature extraction from histopathological images, employing DenseNet and ResNet for feature extracting. The Denoising Sparse AutoEncoder (DSAE) model, improved by the AO algorithm, is employed for BC detection and classification. The trials were carefully carried out with various dataset magnifications, including resolutions at 220px and 400px. Our suggested approaches and techniques could be thoroughly assessed through careful investigation, guaranteeing their resilience and adaptability to different image resolutions.. The HRD-BC2HI technique outperformed other current methods in BC classification when tested on a benchmark dataset. It obtained remarkable results with an average accuracy of 98.92%. The outcomes suggest that the HRD-BC2HI method performs better for BC classification than other contemporary techniques.

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

Empowering Breast Cancer Detection: A Novel Hybrid Transfer Learning Approach with Aquila Optimizer

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Nuzhat Noor Islam Prova

摘要

Breast cancer (BC) is a condition where aberrant breast cells proliferate uncontrollably, resulting in tumors, and the most hazardous disease for women worldwide is BC. It frequently results in severe psychological suffering, including elevated rates of anxiety and depression. Machine learning (ML) and computer-aided diagnosis (CAD) have demonstrated potential in enhancing the identification of BC, specifically in the area of Deep Learning (DL) methods for the classification of histopathological images (HI). These techniques depend on huge amounts of labeled data, which can be challenging to gather. This study presents a new method called Aquila Optimizer (AO) with Hybrid ResNet-DenseNet Enabled Breast Cancer Classification (HRD-BC2HI), combining ResNet and DenseNet models for feature extraction from histopathological images, employing DenseNet and ResNet for feature extracting. The Denoising Sparse AutoEncoder (DSAE) model, improved by the AO algorithm, is employed for BC detection and classification. The trials were carefully carried out with various dataset magnifications, including resolutions at 220px and 400px. Our suggested approaches and techniques could be thoroughly assessed through careful investigation, guaranteeing their resilience and adaptability to different image resolutions.. The HRD-BC2HI technique outperformed other current methods in BC classification when tested on a benchmark dataset. It obtained remarkable results with an average accuracy of 98.92%. The outcomes suggest that the HRD-BC2HI method performs better for BC classification than other contemporary techniques.