摘要
皮肤癌实验中的皮肤性疾病,皮肤癌的早期筛查对于临床干预治疗至关重要.为提高皮肤癌的精准预测,减少发病率和死亡率,本文在残差网络迁移学习模型的基础上,构建一种皮肤癌良恶性判断的计算机辅助诊断网络.实验结果表明,基于残差网络的预测模型,计算机辅助诊断模型的训练集预测精度为84%,测试集预测精度达到90%.基于计算机辅助诊断的皮肤癌良恶性预测诊断模型为临床皮肤癌早期筛查提供借鉴.
Early screening of skin diseases and skin cancer in skin cancer experiments is essential for clinical intervention treatment.In order to improve the accurate prediction of skin cancer and reduce morbidity and mortality,based on the residual network transfer learning model,this paper constructs a computer-aided diagnosis network for the judgment of benign and malignant skin cancer.Experimental results show that the prediction model based on residual network,the training set prediction accuracy of the computer-aided diagnostic model is 84%,and the prediction accuracy of the test set is 81%.The predictive diagnosis model of benign and malignant skin cancer based on computer-aided diagnosis provides a reference for early screening of clinical skin cancer.
作者
杨子勋
陈广新
李长荣
曹文超
YANG Zi-xun;CHEN Guang-xin;LI Chang-rong;CAO Wen-chao(School of Medical Imaging,Mudanjiang Medical University,Heilongjiang Mudanjiang 157011,China;School of Life Sciences,Mudanjiang Medical University,Heilongjiang Mudanjiang 157011,China;School of nursing,Mudanjiang Medical Universiy,Heilongjiang Mudanjiang 157011,China)
出处
《新一代信息技术》
2022年第8期134-138,共5页
New Generation of Information Technology
基金
2021年黑龙江省大学生创新创业训练计划项目(项目编号:202110229028)
关键词
计算机辅助诊断
残差网络
皮肤癌
良恶性
诊断
computer-aided diagnosis
residual network
skin cancer
benign and malignant
diagnosis
