《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (4): 1354-1362.DOI: 10.11772/j.issn.1001-9081.2025040513

• 前沿与综合应用 • 上一篇    

融合改进的ResNet50与集成分类器的皮肤癌分类

秦传东1,2, 索志强1()   

  1. 1.北方民族大学 数学与信息科学学院,银川 750021
    2.宁夏智能信息与大数据处理重点实验室(北方民族大学),银川 750021
  • 收稿日期:2025-05-09 修回日期:2025-07-09 接受日期:2025-07-11 发布日期:2025-07-15 出版日期:2026-04-10
  • 通讯作者: 索志强
  • 作者简介:秦传东(1976—),男,湖北广水人,教授,博士,主要研究方向:智能计算、大数据分析
  • 基金资助:
    宁夏高等教育机构基金资助项目(NYG2024093)

Skin cancer classification integrating improved ResNet50 with ensemble classifier

Chuandong QIN1,2, Zhiqiang SUO1()   

  1. 1.School of Mathematics and Information Sciences,North Minzu University,Yinchuan Ningxia 750021,China
    2.Ningxia Key Laboratory of Intelligent Information and Big Data Processing (North Minzu University),Yinchuan Ningxia 750021,China
  • Received:2025-05-09 Revised:2025-07-09 Accepted:2025-07-11 Online:2025-07-15 Published:2026-04-10
  • Contact: Zhiqiang SUO
  • About author:QIN Chuandong, born in 1976, Ph. D., professor. His research interests include intelligent computing, big data analysis.
  • Supported by:
    Fund of Ningxia Higher Education Institutions(NYG2024093)

摘要:

皮肤癌是全球发病率持续攀升的恶性肿瘤之一,它的早期精准诊断对降低死亡率至关重要。针对现有模型难以满足临床要求及少数类别识别皮肤癌精度低的问题,提出一种融合改进ResNet50与集成分类器的模型。首先,通过灰度黑帽阈值处理和Telea算法去除毛发噪声,再使用合成少数过采样技术(SMOTE)平衡类别分布;其次,采用ResNet50模型提取深层特征,并引入融合空间与通道注意力的软注意力模块聚焦皮肤病变区域;最后,将随机森林、极限梯度提升(XGBoost)、轻量级梯度提升机(LightGBM)、K-近邻(K-NN)和支持向量机(SVM)作为集成分类器,通过软投票法进行集成,进行皮肤癌的早期诊断。在HAM10000、ISIC2019和ISIC2020数据集上的3次独立实验结果表明,所提模型将准确率分别提升到(98.33±0.03)%、(96.15±0.06)%和(99.19±0.02)%,相较于当前主流网络具有更优的特征提取与分类能力,有助于提升早期诊断效果。

关键词: 皮肤癌分类, ResNet50, 合成少数过采样技术, 集成学习, 注意力机制

Abstract:

As a malignant tumor with continuously rising incidence rates globally, skin cancer requires early and precise diagnosis to reduce the mortality rate. To address the challenges of insufficient model performance for clinical requirements and low diagnostic accuracy in minority categories of skin cancers, a model of an improved ResNet50 integrated with an ensemble classifier was proposed. Firstly, hair-induced noise was eliminated through grayscale black-hat threshold processing and Telea algorithm, and Synthetic Minority Over-sampling TEchnique (SMOTE) was used to balance class distribution. Secondly, the deep-level features were extracted using the ResNet50 model, with a soft attention module combining spatial and channel attention mechanisms introduced to focus on skin lesion regions. Finally, an ensemble classifier integrating random forest, eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), K-Nearest Neighbor (K-NN), and Support Vector Machine (SVM) via soft voting was employed, and the proposed model was applied for early diagnosis of skin cancers. The results of three separate experiments on the HAM10000, ISIC2019 and ISIC2020 datasets indicate that the proposed model improves the accuracy to (98.33±0.03)%, (96.15±0.06)% and (99.19±0.02)%, respectively. Compared with current mainstream networks, the proposed model exhibits superior feature extraction and classification capabilities, and is helpful to improve early diagnostic effects.

Key words: skin cancer classification, ResNet50, Synthetic Minority Over-sampling TEchnique (SMOTE), ensemble learning, attention mechanism

中图分类号: