Scholarlink

Analyzing the medical scans plays a crucial role in the diagnosis and prognosis of diseases. The Curated Breast Imaging Subset DDSM (CBIS-DDSM) dataset is a valuable resource that provides insights into breast mammogram scans. Breast Cancer is the foremost reason for ruining the lives of many women worldwide. Early detection and accurate prediction of Breast Cancer can prudent the lives of several innocents. With the recent advancement in the field of deep learning, we have experimented with the usefulness of CNN and Transfer Learning (TL) for better error diagnosis of breast cancer. The CNN model was trained initially with the features extracted by the TL, namely ResNet50, VGG16, and GoogLeNet, to enhance the model’s robustness to predict the new unseen data. We have also analyzed the pros and cons of these three ProgBC-TCNN models over the CNN architecture, and these three ProgBC-TCNN models resulted in a superior accuracy of 96.96%, 98.84%, and 99.81%, respectively. Hence, this simplifies the model’s robustness in detecting breast cancer and makes the proposed model a virtuous model for enhanced prognosis.

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

ProgBC-TCNN: Enhancing Breast Cancer Mammogram Images Prognosis Employing Hybridized CNN with Transfer Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Umang Kumar Agrawal,
  • Bhramara Bar Biswal,
  • Sambit Ranjan Pattanaik,
  • Abhilash Pati,
  • Amrutanshu Panigrahi,
  • Bibhuprasad Sahu

摘要

Analyzing the medical scans plays a crucial role in the diagnosis and prognosis of diseases. The Curated Breast Imaging Subset DDSM (CBIS-DDSM) dataset is a valuable resource that provides insights into breast mammogram scans. Breast Cancer is the foremost reason for ruining the lives of many women worldwide. Early detection and accurate prediction of Breast Cancer can prudent the lives of several innocents. With the recent advancement in the field of deep learning, we have experimented with the usefulness of CNN and Transfer Learning (TL) for better error diagnosis of breast cancer. The CNN model was trained initially with the features extracted by the TL, namely ResNet50, VGG16, and GoogLeNet, to enhance the model’s robustness to predict the new unseen data. We have also analyzed the pros and cons of these three ProgBC-TCNN models over the CNN architecture, and these three ProgBC-TCNN models resulted in a superior accuracy of 96.96%, 98.84%, and 99.81%, respectively. Hence, this simplifies the model’s robustness in detecting breast cancer and makes the proposed model a virtuous model for enhanced prognosis.