Scholarlink

Numerous individuals deal with numerous skin-related issues. Numerous people suffer from different skin conditions, which have historically been a prevalent human difficulty. Conventional approaches to skin condition diagnosis involve numerous tests, are thought to be time-consuming, and necessitate a deep grasp of the field. For the diagnosis, visual evaluation in conjunction with clinical data may be beneficial. Convolutional neural networks (CNNs) and an assembly representation built with VGG16, DenseNet, and Inception are the methods used in this technique. The particular skin disorders that have been taken into account are Actinic Keratoses, Basal Cell Carcinoma, Benign Keratosis, Dermatofibroma, Melanoma, Melanocytic Nevi, and Venous Lesions. The accuracy varies from 71 to 75% when using CNN, 80.3, 82.3, and 80.4% for VGG16, DenseNet, and Inception, respectively, while a mixture of VGG16, DenseNet, and Inception achieves 83–85% accuracy.

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

A Inventive and Practical Method for Increasing the Precision of Machine Learning-Based Skin Disease Detection

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • G. Usharani,
  • M. Prabhakar,
  • S. Rao Chintalapudi,
  • Siddamma,
  • V. Narasimha,
  • Dipika Rath

摘要

Numerous individuals deal with numerous skin-related issues. Numerous people suffer from different skin conditions, which have historically been a prevalent human difficulty. Conventional approaches to skin condition diagnosis involve numerous tests, are thought to be time-consuming, and necessitate a deep grasp of the field. For the diagnosis, visual evaluation in conjunction with clinical data may be beneficial. Convolutional neural networks (CNNs) and an assembly representation built with VGG16, DenseNet, and Inception are the methods used in this technique. The particular skin disorders that have been taken into account are Actinic Keratoses, Basal Cell Carcinoma, Benign Keratosis, Dermatofibroma, Melanoma, Melanocytic Nevi, and Venous Lesions. The accuracy varies from 71 to 75% when using CNN, 80.3, 82.3, and 80.4% for VGG16, DenseNet, and Inception, respectively, while a mixture of VGG16, DenseNet, and Inception achieves 83–85% accuracy.