A sensor-based wrist pulse signal processing and lung cancer recognition

J Biomed Inform. 2018 Mar:79:107-116. doi: 10.1016/j.jbi.2018.01.009. Epub 2018 Feb 8.

Abstract

Pulse diagnosis is an efficient method in traditional Chinese medicine for detecting the health status of a person in a non-invasive and convenient way. Jin's pulse diagnosis (JPD) is a very efficient recent development that is gradually recognized and well validated by the medical community in recent years. However, no acceptable results have been achieved for lung cancer recognition in the field of biomedical signal processing using JPD. More so, there is no standard JPD pulse feature defined with respect to pulse signals. Our work is designed mainly for care giving service conveniently at home to the people having lung cancer by proposing a novel wrist pulse signal processing method, having an insight from JPD. We developed an iterative slide window (ISW) algorithm to segment the de-noised signal into single periods. We analyzed the characteristics of the segmented pulse waveform and for the first time summarized 26 features to classify the pulse waveforms of healthy individuals and lung cancer patients using a cubic support vector machine (CSVM). The result achieved by the proposed method is found to be 78.13% accurate.

Keywords: Cubic support vector machine (CSVM); Feature extraction; Iterative sliding window (ISW); Jin’s pulse diagnosis (JPD); Lung cancer recognition; Pulse signal processing and analysis.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Algorithms
  • Equipment Design
  • Healthy Volunteers
  • Heart Rate
  • Humans
  • Lung Neoplasms / diagnosis*
  • Lung Neoplasms / physiopathology*
  • Medicine, Chinese Traditional
  • Monitoring, Ambulatory / instrumentation*
  • Monitoring, Ambulatory / methods
  • Pattern Recognition, Automated
  • Pulse
  • Signal Processing, Computer-Assisted*
  • Support Vector Machine
  • Time Factors
  • Wrist