Scholarlink

Skin cancer has become a serious public health emergency, with an increasing number of people dying from it. Computer-aided diagnosis methods are increasingly being utilized to help physicians identify skin cancer more precisely and prevent errors brought on by human error. By using machine learning algorithms to classify and identify skin lesions, this study aims to provide a reliable method for detecting skin cancer. The proposed technique categorizes photos utilizing Convolutional Neural Networks (CNNs), a form of deep learning recognized for its efficacy and precision. This research was based on the HAM10000 and ISIC Archive datasets. This study identified skin cancer photos with a pre-trained DenseNet169, Visual Geometry Group 16 (VGG16), and MobileNetV2 architecture. According to the findings, these deep learning models performed well on several standardized datasets, with DenseNet169 achieving 91.2%, VGG16 achieving 88.13%, and MobileNetv2 performing 77.86%. These deep learning models can help dermatologists identify skin cancer more precisely and avoid human error.

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

Evaluating Deep Learning Models for Skin Cancer Detection: Insights from DenseNet169, VGG16, and MobileNetv2

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Anchal Kumari,
  • Punam Rattan

摘要

Skin cancer has become a serious public health emergency, with an increasing number of people dying from it. Computer-aided diagnosis methods are increasingly being utilized to help physicians identify skin cancer more precisely and prevent errors brought on by human error. By using machine learning algorithms to classify and identify skin lesions, this study aims to provide a reliable method for detecting skin cancer. The proposed technique categorizes photos utilizing Convolutional Neural Networks (CNNs), a form of deep learning recognized for its efficacy and precision. This research was based on the HAM10000 and ISIC Archive datasets. This study identified skin cancer photos with a pre-trained DenseNet169, Visual Geometry Group 16 (VGG16), and MobileNetV2 architecture. According to the findings, these deep learning models performed well on several standardized datasets, with DenseNet169 achieving 91.2%, VGG16 achieving 88.13%, and MobileNetv2 performing 77.86%. These deep learning models can help dermatologists identify skin cancer more precisely and avoid human error.