ScholarMate
English简体中文繁體中文
客服热线:400-1616-289
登录注册
期刊论文

A hybrid CNN architecture for skin lesion classification using deep learning

Jasil, S. P. Godlin*; Ulagamuthalvi, V.
SOFT COMPUTING2023-3-22DOI: 10.1007/s00500-023-08035-w
8被引2阅读
引用收藏
分享

摘要

The majority of cases of skin cancer can be identified using a combination of a doctor's ocular inspection, a dermoscopy, and other diagnostic methods. There are a number of issues that make automated skin lesion detection from dermoscopic pictures challenging, including artifacts (hairs), irregularity, lesion form, and irrelevant feature extraction. As a result of issues like these, the segmentation and classification procedure are more challenging than they should be. Several skin lesion categorization approaches using deep learning based on convolution neural network (CNN) and annotated skin photos show enhanced outcomes because visual observation provides the possibility to utilize artificial intelligence to intercept various skin images. Most current deep learning approaches to skin lesion segmentation derive their predictions from an assembly of various convolutional neural networks (CNN), the aggregation of multi-scale information, or a multi-task learning framework. The key goal is to utilize as much data as possible for accurate forecasting. The skin lesion segmentation task, which is typically paired with the skin lesion classification task, has been shown to benefit from a multi-task learning framework. In this work, we present a new convolutional neural network (CNN) architecture called Densenet and residual network that makes use of contextual data. To test how well our model performed, we looked at examples from the official Human Against Machine dataset (HAM10000), a collection of images from multiple sources. In order to increase the classifier's effectiveness, we up-sampled the data and supplemented it with additional information. The experimental results show that the approaches proposed here enhance the automatic classification of skin lesions with an accuracy of 95%.

学科与关键词

关键词
ArchitectureConvolutional neural networkComputer-aided diagnosisDeep learningSkin lesion
相关
加载中…
  • SkinCancerNet: Automated Classification of Skin Lesion Using Deep Transfer Learning Method
    Tasar, Beyda*Traitement du Signal, 2023, 40(1): 285-295.被引 3
  • Recent Advances in Diagnosis of Skin Lesions Using Dermoscopic Images Based on Deep Learning
    Nie, Yali*; Sommella, Paolo; Carratu, Marco; Ferro, Matteo; O'Nils, Mattias; Lundgren, Jan*IEEE Access, 2022, 10: 95716-95747.被引 7
  • Skin lesion detection using an ensemble of deep models: SLDED
    Shahsavari, Ali; Khatibi, Toktam*; Ranjbari, SimaMULTIMEDIA TOOLS AND APPLICATIONS, 2023, 82(7): 10575-10594.被引 12
查看全部 →

发表信息

论文状态
公开发表
期刊
SOFT COMPUTING
发表日期
2023-03-22
DOI
10.1007/s00500-023-08035-w
语言
ENGLISH
成果收录
SCIE

作者

  • J

    Jasil, S. P. Godlin (通讯) (第一作者)

  • U

    Ulagamuthalvi, V.

登录解锁更多功能

  • 收藏这篇,随时在“我的 · 收藏”找回
  • 认领你的成果,生成推广素材
  • 关注作者获取最新成果
  • 加入项目组、协同写作
注册 / 登录

产品服务

  • 科研之友
  • 创新城
  • 科创云

服务支持

  • 帮助中心
  • 隐私政策
  • 服务条款

联系方式

在线客服:【】
客服热线:400-1616-289
电子邮箱:support@scholarmate.com

关注或下载科研之友

微信二维码
微信公众号
客户端下载二维码
下载客户端
科研成果科研人员科研机构科研动态爱瑞思软件

©2026 深圳市科研之友网络服务有限公司

公安备案图标粤公网安备 44030502000213
粤ICP备 16046710 号粤B2-20110417