Scholarlink

<p>In this paper, a new system for categorising and dividing breast disease images is provided. The Curated Breast Imaging Subset of DDSM (CBIS-DDSM) is divided into innocuous and harmful using a variety of models, including the VGG16, VGG19, and ResNet50 models. Additionally, the breast region from the mammography images is fragmented using the provided changed U-Net model. This tactic will assist a radiologist in making early discoveries and improve the effectiveness of our framework. The identification and diagnosis of breast disease typically involve the use of the Cranio Caudal (CC) vision and Mediolateral Oblique (MLO) views. To improve the framework execution, our suggested outline work rely on MLO view and CC view. Additionally, the lack of labelled information is a significant test. Move learning and information increase are applied to conquer this issue. This accomplishes 96% accuracy, 94.66% F1 score on DDSM datasets.</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

RETRACTED ARTICLE: Detection and classification of mammogram using ResNet-50

  • 期刊论文

  • 发表日期：

    2025年2月28日

查看全文
  • Rupali A. Patil,
  • V. V. Dixit

摘要

In this paper, a new system for categorising and dividing breast disease images is provided. The Curated Breast Imaging Subset of DDSM (CBIS-DDSM) is divided into innocuous and harmful using a variety of models, including the VGG16, VGG19, and ResNet50 models. Additionally, the breast region from the mammography images is fragmented using the provided changed U-Net model. This tactic will assist a radiologist in making early discoveries and improve the effectiveness of our framework. The identification and diagnosis of breast disease typically involve the use of the Cranio Caudal (CC) vision and Mediolateral Oblique (MLO) views. To improve the framework execution, our suggested outline work rely on MLO view and CC view. Additionally, the lack of labelled information is a significant test. Move learning and information increase are applied to conquer this issue. This accomplishes 96% accuracy, 94.66% F1 score on DDSM datasets.