Scholarlink

This study investigates applying convolutional neural networks (CNNs) for real-time analysis of cold atmospheric plasma, introducing a novel architecture with inception modules, residual blocks, and attention mechanisms to improve feature extraction and accuracy. The model is trained on synthetic plasma diagnostic data, showing a decrease in training mean absolute error (MAE) from 0.255 to 0.175 over 10 epochs, and a slight decrease in validation MAE from 0.295 to 0.275, indicating effective learning and generalization. The proposed CNN achieved a Test MAE of 0.27 and Test Loss of 0.10.

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

Deep Learning for Real-Time Diagnostics of Cold Atmospheric Plasma

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Raghavendra M. Devadas,
  • Preethi,
  • R. Sapna

摘要

This study investigates applying convolutional neural networks (CNNs) for real-time analysis of cold atmospheric plasma, introducing a novel architecture with inception modules, residual blocks, and attention mechanisms to improve feature extraction and accuracy. The model is trained on synthetic plasma diagnostic data, showing a decrease in training mean absolute error (MAE) from 0.255 to 0.175 over 10 epochs, and a slight decrease in validation MAE from 0.295 to 0.275, indicating effective learning and generalization. The proposed CNN achieved a Test MAE of 0.27 and Test Loss of 0.10.