Scholarlink

<p>Active noise cancellation remains a complex and intriguing problem for the design of hearing aids, telecommunication devices, and teleconferencing platforms. Many scholars have developed algorithms for improving voice signal quality and reducing noise in the past decades. The objective of this research is to reduce noise while improving the signal-to-noise ratio (SNR) of noisy speech transmissions in an unfavourable environment. In this research, a complex Multitude active noise cancellation via Fennec Fox Optimized CNN-BiLSTM Network (CMANC-Net) has been proposed. Initially, Fast Independent Component Analysis (F-ICA) is utilized to enhances the quality of the signal by leveraging the statistical independence of complex multitude noise. Community based Genetic bee colony optimization is utilized to extract the most relevant feature from the filtered signal. Moreover, hybrid Convolutional Neural Network based BiLSTM is utilized for classifying desired and interference signal. Also, the Fennec Fox Optimization method is used for tuning the hyper parameters of the neural network to increase classification accuracy. To assess the efficacy of proposed approaches, performance measurements like STOI, NMSE, specificity, accuracy, sensitivity, and PESQ are employed. The proposed method achieves an overall classification accuracy of 99.37% for the Urban 8k dataset and 99.05% for the NoiseX-92 dataset.</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

CMANC Net: Fennec Fox Optimized CNN-BiLSTM Network for Real Time and Complex Multitude Active Noise Cancellation

  • 期刊论文

  • 发表日期：

    2025年10月5日

查看全文
  • V. D. M. Jabez,
  • A. Ahilan,
  • Chukka Santhaiah,
  • Vignesh Thangathurai

摘要

Active noise cancellation remains a complex and intriguing problem for the design of hearing aids, telecommunication devices, and teleconferencing platforms. Many scholars have developed algorithms for improving voice signal quality and reducing noise in the past decades. The objective of this research is to reduce noise while improving the signal-to-noise ratio (SNR) of noisy speech transmissions in an unfavourable environment. In this research, a complex Multitude active noise cancellation via Fennec Fox Optimized CNN-BiLSTM Network (CMANC-Net) has been proposed. Initially, Fast Independent Component Analysis (F-ICA) is utilized to enhances the quality of the signal by leveraging the statistical independence of complex multitude noise. Community based Genetic bee colony optimization is utilized to extract the most relevant feature from the filtered signal. Moreover, hybrid Convolutional Neural Network based BiLSTM is utilized for classifying desired and interference signal. Also, the Fennec Fox Optimization method is used for tuning the hyper parameters of the neural network to increase classification accuracy. To assess the efficacy of proposed approaches, performance measurements like STOI, NMSE, specificity, accuracy, sensitivity, and PESQ are employed. The proposed method achieves an overall classification accuracy of 99.37% for the Urban 8k dataset and 99.05% for the NoiseX-92 dataset.