System and method of detecting abnormal movement of a physical object

Patent metadata

Field	Value
Publication number	US-9314177-B2
Application number	US-201314012873-A
Country	US
Kind code	B2
Filing date	Aug 28, 2013
Priority date	Dec 28, 2010
Publication date	Apr 19, 2016
Grant date	Apr 19, 2016

Abstract

The present application discloses a method of detecting abnormal movement of a physical object. A periodic signal is representative of the movement of the object. According to some embodiments, a raw matrix having a first array and a second array is generated, and then an integrated matrix is generated by performing a dimension reduction on the raw matrix. A likelihood of a predetermined type of abnormal movement of the physical object is determined by comparing the integrated matrix with a predetermined benchmark pattern. In some embodiments, the generation of the raw matrix includes performing a first analysis on a predetermined portion of the periodic signal to generate the first array and performing a second analysis different from the first analysis on the predetermined portion of the periodic signal to generate the second array.

First claim

What is claimed is:

1. A method of detecting abnormal movement of a physical object, the method comprising:
generating, using a transducer, a periodic signal representative of a movement of the physical object,
generating, by a processor, a raw matrix comprising a first array and a second array, the generation of the raw matrix comprising:
performing a first analysis on a predetermined portion of the periodic signal to generate the first array, the predetermined portion corresponding to a predetermined time period of the periodic signal; and
performing a second analysis different from the first analysis on the predetermined portion of the periodic signal to generate the second array;

generating an integrated matrix by performing a dimension reduction on the raw matrix; and
determining a likelihood of a predetermined type of abnormal movement of the physical object by comparing the integrated matrix or a set of indexes derived from the integrated matrix with a predetermined benchmark pattern corresponding to the predetermined type of abnormal movement,
wherein the first analysis is a mutual information analysis, comprising:
segmenting the predetermined duration of the periodic signal into a plurality of signal segments, each signal segment including a nominal waveform, and the plurality of signal segments including one or more pairs of adjacent signal segments;
for each pair of the one or more pairs of adjacent signal segments:
obtaining two signal arrays having the same size by re-sampling one or both of the adjacent signal segments; and
calculating joint probability and marginal probability of the two signal arrays of the re-sampled adjacent signal segments;

calculating a mutual information index based on the calculated joint probability and the calculated marginal probability of the one or more pairs of re-sampled adjacent signal segments; and
outputting the likelihood of the predetermined type of abnormal movement of the physical object to a display.

2. The method of claim 1 , wherein the second analysis is a time-domain analysis, a pattern analysis, or deriving a feature from the predetermined duration of the periodic signal obtained according to a first spatial measurement configuration and at least a portion of another periodic signal being representative of the movement obtained according to a second spatial measurement configuration.

3. The method of claim 1 , wherein the calculation of the mutual information index is performed based on an equation of:

M
⁢

⁢

I
⁡

(

X
;
Y

)

=

∑
x

⁢

∑
y

⁢

p
⁡

(

x
,
y

)

⁢
log
⁢

p
⁡

(

x
,
y

)

p
⁡

(
x
)

⁢

p
⁡

(
y
)

where x and y each represents components of one of the two signal arrays of the re-sampled adjacent signal segments, p(x,y) represents the joint probability of the two signal arrays, and p(x) and p(y) each represents the marginal probability of one of the two signal arrays of the re-sampled adjacent signal segments.

4. The method of claim 1 , wherein the transducer is an electrocardiography (ECG) transducer.

5. The method of claim 4 , wherein the second analysis is a time-domain analysis, a morphology analysis, or an electric-axis analysis.

6. The method of claim 4 , wherein the predetermined type of abnormal movement corresponds to a predetermined illness.

7. The method of claim 1 , wherein the dimension reduction is performed by applying principal component analysis, factor analysis, or independent component analysis on the raw matrix.

8. The method of claim 1 , wherein the raw matrix further comprises a third array, and the generation of the raw matrix further comprises performing a third analysis different from the first analysis and the second analysis on the predetermined portion of the periodic signal to generate the third array.

9. The method of claim 1 , wherein the periodic signal is obtained by detecting one or more of the following activities:
heartbeat, breathing, speech, earthquake or seismic activity, orbital movement of an astronomical object, periodic variances of an astronomical object, movement of a piston, or rotation of a motor.

10. The method of claim 1 , wherein at least one pair of the one more pairs of adjacent signal segments have different lengths.

Assignees

• Ind Tech Res Inst

Inventors

• Ting Chuan-Wei
• Wang Ching-Yao

Classifications

• A61B5/349 Primary

  Detecting specific parameters of the electrocardiograph cycle · CPC title

• A61B5/0452

  Human Necessities · mapped topic

• A61B5/04012 Primary

  Human Necessities · mapped topic

• A61B5/0245

  by using sensing means generating electric signals, {i.e. ECG signals} · CPC title

• A61B5/0816

  Measuring devices for examining respiratory frequency (measuring frequency of electric signals G01R23/00) · CPC title