Systems and methods for detecting anomaly in a cardiovascular signal using hierarchical extremas and repetitions

Patent metadata

Field	Value
Publication number	US-11304663-B2
Application number	US-201816230053-A
Country	US
Kind code	B2
Filing date	Dec 21, 2018
Priority date	Dec 29, 2017
Publication date	Apr 19, 2022
Grant date	Apr 19, 2022

Abstract

Systems and methods for detecting an anomaly in a cardiovascular signal using hierarchical extremas and repetitions. The traditional systems and methods provide for some anomaly detection in the cardiovascular signal but do not consider the discrete nature and strict rising and falling patterns of the cardiovascular signal and frequency in terms of hierarchical maxima points and minima points. Embodiments of the present disclosure provide for detecting the anomaly in the cardiovascular signal using hierarchical extremas and repetitions by smoothening the cardiovascular signal, deriving sets of hierarchical extremas using window detection, identifying signal patterns based upon the sets of hierarchical extremas, identifying repetitions in the signal patterns based upon occurrences and randomness of occurrences of the signal patterns and classifying the cardiovascular signal as anomalous and non-anomalous for detecting the anomaly in the cardiovascular signal.

First claim

What is claimed is:

1. A method for detecting an anomaly in a cardiovascular signal using hierarchical extremas and repetitions, the method comprising a processor implemented steps of:
smoothening, using a filter, the cardiovascular signal acquired for filtering the cardiovascular signal;
deriving, using a window detection technique, one or more sets of hierarchical extremas, based upon the smoothened cardiovascular signal, wherein the one or more sets of hierarchical extremas comprises maximum points and minimum points based on rising edges and falling edges of the cardiovascular signal, wherein each level of hierarchy in the one or more sets of hierarchical extremas represents a different window size of detection, and wherein the window detection technique to derive the one or more sets of hierarchical extremas comprising the steps of:
deriving a number of sampling points and amplitude difference from minima to maxima and maxima to minima;
performing clustering on the number of sampling points and the amplitude difference from the minima to maxima using k-means clustering to derive a number of clusters with centroids;
performing clustering on the number of sampling points and the amplitude difference from the maxima to minima using the k-means clustering to derive a number of clusters with centroids;
determining a plurality of boundaries of the number of clusters for the minima to maxima;
determining a plurality of boundaries of the number of clusters for the maxima to minima;
obtaining a value of window left (w l ) as a function of the plurality of boundaries of the number of clusters corresponding to the number of sampling points and the amplitude difference from the minima to maxima;
obtaining a value of window right (w r ) as a function of the plurality of boundaries of the number of clusters corresponding to the number of sampling points and the amplitude difference from the maxima to minima; and
obtaining a primary window (T p ), a secondary window (T st ) and a tertiary window (T t ) using the obtained value of w l and w r , for deriving the one or more sets of hierarchical extremas, wherein the Vis represented as T p =w l +w r , the T st is represented as T st =w l /2 and the T t is represented as T t =(w l +w r )/4;

identifying, one or more elements of signal patterns, based upon the one or more sets of hierarchical extremas, wherein the one or more elements of signal patterns comprise multiple frequencies and significance associated with the cardiovascular signal for defining a plurality of physiological events of the user or noise, wherein identifying the significance of the one or more elements of the signal patterns comprises obtaining a lower triangular matrix based upon the one or more sets of hierarchy of extremas, and wherein the lower triangular matrix comprises number of occurrences of the one or more elements of signal patterns to identify variability in the cardiovascular signal;
detecting, the anomaly in the cardiovascular signal by:
determining occurrences of the one or more elements of signal patterns;
determining randomness of occurrences of the one or more elements of signal patterns, by computing an entropy of occurrences of the one or more elements of signal patterns, wherein the entropy comprises randomness of the one or more elements of signal patterns computed based upon probabilities of repetitions of the one or more elements of signal patterns, and wherein determining the randomness of occurrences of the one or more elements of signal patterns comprises obtaining one or more threshold values based upon an equi-probable occurrence of the one or more elements of signal patterns for classifying the one or more elements of signal patterns; and
identifying, significance of repetitions of the one or more elements of signal patterns, based upon the occurrences and randomness of occurrences to detect the anomaly in the cardiovascular signal, wherein identifying the significance of repetitions of the one or more elements of signal patterns comprises obtaining a lower triangular matrix based upon the one or more sets of hierarchy of extremas, and wherein the lower triangular matrix comprises number of occurrences of the one or more elements of signal patterns to identify variability in the cardiovascular signal.

2. The method of claim 1 , wherein the step of identifying the significance of the one or more elements of signal patterns further comprises evaluating entropy of elements of a lower triangular matrix based upon frequencies and number of points in the one or more elements of signal patterns to detect randomness of the one or more elements of signal patterns.

3. The method of claim 1 , wherein the step of identifying the one or more elements of signal patterns is preceded by:
(i) detecting, one or more zero patterns in the cardiovascular signal based upon the one or more sets of hierarchical extremas; and
(ii) filtering, the one or more zero patterns, based upon a comparison of the one or more zero patterns and a predefined threshold to detect the anomaly in the cardiovascular signal.

4. The method of claim 1 , wherein the step of identifying the one or more elements of signal patterns further comprises identifying uni-modal and multi-modal patterns in the cardiovascular signal based upon the occurrences of the one or more elements of signal patterns to detect the anomaly.

5. The method of claim 1 , wherein the step of obtaining the one or more threshold values comprises computing an upper threshold value based upon occurrences and henceforth entropy of the one or more elements of signal patterns to detect the anomaly.

6. A system for detecting an anomaly in a cardiovascular signal using hierarchical extremas and repetitions, the said system comprising:
a memory storing instructions;
one or more communication interfaces; and
one or more hardware processors coupled to the memory via the one or more communication interfaces, wherein the one or more hardware processors are configured by the instructions to:
smoothen, using a filter, the cardiovascular signal acquired for filtering the cardiovascular signal;
derive, using a window detection technique, one or more sets of hierarchical extremas, based upon the smoothened cardiovascular signal, wherein the one or more sets of hierarchical extremas comprises maximum points and minimum points based on rising edges and falling edges of the cardiovascular signal, wherein each level of hierarchy in the one or more sets of hierarchical extremas represents a different window size of detection, and wherein the window detection technique to derive the one or more sets of hierarchical extremas comprising the steps of:
deriving a number of sampling points and amplitude difference from minima to maxima and maxima to minima;
performing clustering on the number of sampling points and the amplitude difference from the minima to maxima using k-means clustering to derive a number of clusters with centroids;
performing clustering on the number of sampling points and the amplitude difference from the maxima to minima using the k-means clustering to derive a number of clusters with centroids;
determining a plurality of boundaries of the number of clusters for the minima to maxima;
determining a plurality of boundaries of the number of clusters for the maxima to minima;
obtaining a value of window left (w l ) as a function of the plurality of boundaries of the number of clusters corresponding to the number of sampling points and the amplitude difference from the minima to maxima;
obtaining a value of window right (w r ) as a function of the plurality of boundaries of the number of clusters corresponding to the number of sampling points and the amplitude difference from the maxima to minima; and
obtaining a primary w

Assignees

• Tata Consultancy Services Ltd

Inventors

• Bandyopadhyay Soma
• Ukil Arijit
• Puri Chetanya
• Singh Rituraj
• Pal Arpan
• Murthy C A

Classifications

• G06F2218/02

  Preprocessing · CPC title

• G16H50/30 Primary

  for calculating health indices; for individual health risk assessment · CPC title

• A61B5/02416

  using photoplethysmograph signals, e.g. generated by infrared radiation (A61B5/14552 takes precedence) · CPC title

• G16H50/70

  for mining of medical data, e.g. analysing previous cases of other patients · CPC title

• G16H50/20

  for computer-aided diagnosis, e.g. based on medical expert systems · CPC title