Scholarlink

<p>Electrocardiogram (ECG) signals in smart health care systems are essential for diagnosing cardiac abnormalities, but their large size poses a challenge for storage and transmission. This requires effective data compression along with efficient transmission. This work offers a systematic way for compressing cardiac signals to improve transmission efficiency. The approach is of two steps using complementary ensemble empirical mode decomposition (CEEMD) and LSTM-based autoencoder (LSTM-AE). The experiment uses the Physionet database&#xa0;that is recorded by the numbers&#xa0;100 and 202. Mendeley Database record 202 is examined and tested to assure model efficiency. To identify the essential features of the ECG signal, the signal is first decomposed and then denoised using CEEMD in a single step. In addition, the signals that have been decomposed are compressed and encoded using the LSTM-AE model. The proposed method achieves a compression ratio (CR) of 38.26 with a percent root mean square difference (PRD) of 0.37 for the Mendeley database record 202, and a CR of 38.25 with a PRD of 0.31 for the Physionet database record 202. LSTM-AE results in a more accurate representation of the signal at any given moment and maintains a balanced resolution throughout.</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

An Ensemble Technique for Cardiac Data Compression in Smart Healthcare System

  • 期刊论文

  • 发表日期：

    2025年1月15日

查看全文
  • Mihir Narayan Mohanty,
  • Sudeshna Baliarsingh,
  • Prakash Kumar Panda

摘要

Electrocardiogram (ECG) signals in smart health care systems are essential for diagnosing cardiac abnormalities, but their large size poses a challenge for storage and transmission. This requires effective data compression along with efficient transmission. This work offers a systematic way for compressing cardiac signals to improve transmission efficiency. The approach is of two steps using complementary ensemble empirical mode decomposition (CEEMD) and LSTM-based autoencoder (LSTM-AE). The experiment uses the Physionet database that is recorded by the numbers 100 and 202. Mendeley Database record 202 is examined and tested to assure model efficiency. To identify the essential features of the ECG signal, the signal is first decomposed and then denoised using CEEMD in a single step. In addition, the signals that have been decomposed are compressed and encoded using the LSTM-AE model. The proposed method achieves a compression ratio (CR) of 38.26 with a percent root mean square difference (PRD) of 0.37 for the Mendeley database record 202, and a CR of 38.25 with a PRD of 0.31 for the Physionet database record 202. LSTM-AE results in a more accurate representation of the signal at any given moment and maintains a balanced resolution throughout.