Scholarlink

<p>Facial emotion classification is very important in health and education. This paper adapts a model that integrates CNNs, ResNet, and an XGBoost meta-model to improve classification precision in both of them. The outputs from both of them are fed into an XGBoost meta-model to make robust classifications. With an accuracy rate of 99.39% on AffectNet and 99.42% on the Facial Emotion Recognition Image dataset, this model surpasses AdaBoost and logistic regression, showing excellent efficiency and robustness for applications in healthcare, education, and security.</p>

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 Stacked CNN-ResNet-XGBoost Framework for Enhanced Facial Emotion Classification

  • 期刊论文

  • 发表日期：

    2025年9月11日

查看全文
  • Ayesha Shaik,
  • Jincy Jis Kanichai,
  • Laxmi Sasikumar,
  • R. Karthik,
  • A. Balasundaram

摘要

Facial emotion classification is very important in health and education. This paper adapts a model that integrates CNNs, ResNet, and an XGBoost meta-model to improve classification precision in both of them. The outputs from both of them are fed into an XGBoost meta-model to make robust classifications. With an accuracy rate of 99.39% on AffectNet and 99.42% on the Facial Emotion Recognition Image dataset, this model surpasses AdaBoost and logistic regression, showing excellent efficiency and robustness for applications in healthcare, education, and security.