Data processing method for respiratory detection based on temperature difference method by deep learning

By using deep learning to adaptively adjust the sampling frequency and construct a stepped noise diagnostic model, the problems of signal interference and rigid sampling strategies in the temperature difference method of respiratory detection are solved, and efficient and accurate respiratory signal acquisition and analysis are achieved.

CN122271997APending Publication Date: 2026-06-26SUN YAT SEN MEMORIAL HOSPITAL SUN YAT SEN UNIV
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
SUN YAT SEN MEMORIAL HOSPITAL SUN YAT SEN UNIV
Filing Date
2026-04-20
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Existing temperature difference-based respiration detection technology faces problems such as signal susceptibility to environmental coupling interference, rigid sampling strategies, and insufficient targeting of traditional denoising methods, resulting in poor robustness, low acquisition efficiency, and lack of adaptability.

Method used

By using a deep learning-based approach, the sampling frequency is adaptively adjusted to construct a stepped noise diagnostic model, directional threshold correction and median smoothing are performed, and a closed-loop feedback mechanism for dual-dimensional weighted evaluation is established to trigger the physical correction of the sensor's spatial isolation distance.

Benefits of technology

It significantly improves the system's anti-interference capability and robustness against complex environmental noise, ensuring high-quality acquisition and accurate analysis of respiratory signals, and improving data acquisition efficiency and accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122271997A_ABST
    Figure CN122271997A_ABST
Patent Text Reader

Abstract

This invention relates to the field of respiratory detection data processing technology, and more particularly to a data processing method for respiratory detection using the temperature difference method based on deep learning. The method includes steps S1, adaptive sampling frequency adjustment; S2, respiratory feature vector extraction; S3, stepped noise diagnosis; S4, deep model adaptive denoising; and S5, closed-loop feedback optimization. This invention ensures accurate capture of transient respiratory signals by dynamically adjusting the sampling frequency according to signal acceleration through nonlinear stepped adaptive sampling; it significantly suppresses environmental thermal coupling noise by performing dynamic threshold correction denoising through a deep learning model; and it dynamically optimizes the model learning rate or sensor spatial isolation distance using dual-dimensional weighted evaluation results, achieving closed-loop feedback optimization of software parameters and hardware configuration. This method effectively improves the robustness, anti-interference ability, and real-time performance of temperature difference method respiratory signal acquisition.
Need to check novelty before this filing date? Find Prior Art