Scholarlink

<p>In recent years, emotion recognition has involved growing interest because of its broad range of applications across areas like healthcare, education, and human-computer interaction. Moreover, analyzing and understanding facial expressions helps to improve the interaction between humans and machines. However, traditional methods often struggle to accurately recognize emotions and make decisions based on them, which is vital in systems that deal with human well-being and safety. Therefore, this paper introduces a Coyote and Badger Makeup Artist Optimization-based Convolutional Deep High-order Attention Neural network (CBMAO_CHA-Net) for sentiment emotion classification based on multimodal data. Firstly, the input Electroencephalography (EEG) signal is pre-processed based on a Gaussian filter. Next, key features are identified through the process of feature extraction. Simultaneously, frame extraction is performed on the input video. Next, face detection is carried out based on Region-based Convolutional Neural Network (RCNN), and feature extraction is done to achieve Locally Adaptive Regression Kernel (LARK) and statistical measures. Then, the extracted features from the EEG signal and video are utilized for sentiment emotion classification, and it is done using CHA–Net, a hybrid model combining a Convolutional Neural Network (CNN) and a Deep High-order Attention Neural Network (DHA–Net). Moreover, the hyperparameters of CHA–Net are tuned using CBMAO, which is developed by combining Coyote and Badger Optimization (CBO) and Makeup Artist Optimization (MAOA). Moreover, the proposed CBMAO_CHA-Net outperforms existing methods by achieving an accuracy of 97.55%, a Sensitivity of 96.94%, and a Specificity of 97.58%.</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

Coyote and Badger Makeup Artist Optimization based Hybrid Deep Learning for Multimodal Sentiment Classification with Emotion Recognition

  • 期刊论文

  • 发表日期：

    2026年3月24日

查看全文
  • T. V. Neethu,
  • E. Grace Mary Kanaga

摘要

In recent years, emotion recognition has involved growing interest because of its broad range of applications across areas like healthcare, education, and human-computer interaction. Moreover, analyzing and understanding facial expressions helps to improve the interaction between humans and machines. However, traditional methods often struggle to accurately recognize emotions and make decisions based on them, which is vital in systems that deal with human well-being and safety. Therefore, this paper introduces a Coyote and Badger Makeup Artist Optimization-based Convolutional Deep High-order Attention Neural network (CBMAO_CHA-Net) for sentiment emotion classification based on multimodal data. Firstly, the input Electroencephalography (EEG) signal is pre-processed based on a Gaussian filter. Next, key features are identified through the process of feature extraction. Simultaneously, frame extraction is performed on the input video. Next, face detection is carried out based on Region-based Convolutional Neural Network (RCNN), and feature extraction is done to achieve Locally Adaptive Regression Kernel (LARK) and statistical measures. Then, the extracted features from the EEG signal and video are utilized for sentiment emotion classification, and it is done using CHA–Net, a hybrid model combining a Convolutional Neural Network (CNN) and a Deep High-order Attention Neural Network (DHA–Net). Moreover, the hyperparameters of CHA–Net are tuned using CBMAO, which is developed by combining Coyote and Badger Optimization (CBO) and Makeup Artist Optimization (MAOA). Moreover, the proposed CBMAO_CHA-Net outperforms existing methods by achieving an accuracy of 97.55%, a Sensitivity of 96.94%, and a Specificity of 97.58%.