Combination non-invasive and invasive bioparameter measuring device
Inventors
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-9844345-B2
Publication Date
2017-12-19
Expiration Date
Abstract
In a combination invasive and non-invasive bioparameter monitoring device an invasive component measures the bioparameter and transmits the reading to the non-invasive component. The non-invasive component generates a bioparametric reading upon insertion by the patient of a body part. A digital processor processes a series over time of digital color images of the body part and represents the digital images as a signal over time that is converted to a learning vector using mathematical functions. A learning matrix is created. A coefficient of learning vector is deduced. From a new vector from non-invasive measurements, a new matrix of same size and structure is created. Using the coefficient of learning vector, a recognition matrix may be tested to measure the bioparameter non-invasively. The learning matrix may be expanded and kept regular. After a device is calibrated to the individual patient, universal calibration can be generated from sending data over the Internet.
Core Innovation
The invention concerns monitoring a bioparameter by invasively measuring the bioparameter of a patient using an invasive component and transmitting and storing an invasive bioparameter reading in a non-invasive component. Within a proximity time of the invasive measuring, one or more variable sensors in the non-invasive component generate a series of data about tissue of a body part and convert sensed magnitudes into electric signals representing a non-invasive measurement of the bioparameter.
The non-invasive component uses one or more processors to convert the signal into a scalar learning number by a mathematical function and repeats this conversion to form learning vectors corresponding to entries of a column vector Y. From a plurality of learning vectors, one or more processors form an n by n learning matrix D that is a regular matrix, repeating invasively and non-invasively measuring enough times to obtain sufficient correlations between non-invasive bioparametric readings and invasive bioparametric readings so that the bioparameter can be measured using a non-invasive bioparameter reading at a pre-defined level of threshold acceptability.
For a new non-invasive measurement, the processors obtain a new vector Vnew by converting signals from a new series of data into scalar numbers. The invention either forms a regular matrix Dnew of n by n size with a non-zero element structure identical to learning matrix D and computes coefficients of a learning vector, or instead compares Vnew with rows of learning matrix D to identify a best match.
In the matrix-vector approach, the processors obtain a coefficient of learning vector C by multiplying an inverse matrix D−1 of learning matrix D by the column vector Y, then perform a matrix vector multiplication of Dnew by coefficient of learning vector C to create a column vector of non-invasive bioparameter measurement R. The processors compare entries of R with entries of Y to find a calibrated bioparameter value for the patient. In the best-match approach, the processors use Vnew to compare with rows of learning matrix D, and use an ith entry of Y as a calibrated bioparameter value for the patient.
Claims Coverage
The provided set includes multiple independent claims. The independent claims cover four inventive features: paired invasive/non-invasive measurement with learning-matrix calibration, a portable dual-component bioparameter-monitoring device, patient-tailored production by calibration under a predefined standard, and software implementing the learning-vector/matrix algorithms to output calibrated bioparameter values.
Learning-vector conversion into a regular n×n learning matrix for correlated acceptability
Convert non-invasive sensor signals into scalar learning numbers using mathematical functions and repeat to form learning vectors corresponding to invasive entries in a column vector Y, then form an n by n learning matrix D that is a regular matrix by repeating invasively measuring and non-invasively generating paired data enough times that sufficient correlations are obtained to measure the bioparameter using a non-invasive bioparameter reading at a pre-defined level of acceptability.
Calibrating a new non-invasive measurement using Vnew and Dnew with identical non-zero element structure
Obtain a new vector Vnew by converting new non-invasive signals into scalar numbers, then use entries of Vnew to form a regular matrix Dnew of n by n size whose structure of non-zero elements is identical to a structure of non-zero elements of learning matrix D, and use a coefficient of learning vector C obtained by multiplying an inverse matrix D−1 by the column vector Y to perform a matrix vector multiplication of Dnew by C to create a column vector R and compare entries of R with entries of Y to find a calibrated bioparameter value for the patient.
Calibrating a new non-invasive measurement by best-match comparison of Vnew with rows of D
Use one or more digital processors to compare entries of Vnew with rows of learning matrix D to find a best match involving an ith row of learning matrix D, and use the ith entry of Y as a calibrated bioparameter value for the patient.
Proximity time paired invasive/non-invasive measurement with stored invasive reading in non-invasive component
Invasively measure the bioparameter of a patient using an invasive component, transmit the invasive bioparameter reading to a non-invasive component and store the invasive bioparameter reading in the non-invasive component, and within a proximity time of the invasive measuring generate non-invasive sensor data and convert sensed magnitudes into electric signals representing a non-invasive measurement of the bioparameter.
Portable dual-component bioparameter-monitoring device with invasive-to-non-invasive digital processing and calibration
Provide a non-invasive component configured to generate non-invasive bioparametric readings of tissue over time using at least one variable sensor and a first digital processor, and provide an invasive component for obtaining an invasive bioparametric reading from blood with automatic transmission to the first digital processor, with the first digital processor programmed to convert signals to scalar learning numbers and form a learning vector corresponding to a scalar invasive bioparameter reading entry of a column vector Y, form a regular n by n learning matrix D by repeating sufficient paired measurements for threshold acceptability, generate Vnew, and either compute calibrated values via Dnew and inverse-matrix coefficients or via best-match comparison of Vnew with rows of D.
Patient-tailored production by calibrating the non-invasive component to approximate invasive readings under a predefined standard
Produce a portable bioparameter-monitoring medical device custom-tailored to a patient by providing a non-invasive component and an invasive component operatively engaged via a coupling element for transmission of invasive bioparametric readings to a first digital processor, and calibrating the non-invasive component to the patient by invasively measuring and transmitting invasive readings and non-invasively measuring within proximity time enough times to obtain sufficient correlations so that the non-invasive bioparametric readings approximate the invasive bioparametric readings for a given bioparameter under a predefined standard of approximation at a pre-defined level of acceptability.
Software implementing scalar learning number conversion, learning matrices, and calibrated output by comparing R and Y
Execute bioparameter monitoring software that converts a signal generated by a non-invasive component into a scalar learning number using mathematical functions, form learning vectors corresponding to invasive entries in a column vector Y, form an n by n regular learning matrix D from a plurality of learning vectors, obtain a new vector Vnew, use entries of Vnew to form a regular matrix Dnew with identical non-zero element structure to learning matrix D, obtain a coefficient of learning vector C by multiplying an inverse matrix D−1 by Y, perform matrix vector multiplication of Dnew by C to create a column vector R, and compare entries of R with entries of Y to find a calibrated bioparameter value for the patient.
Across the independent claims, the inventive coverage centers on converting non-invasive signals into scalar learning numbers to build a regular learning matrix D from paired invasive/non-invasive readings at a proximity-time standard and predefined acceptability threshold, then calibrating a new non-invasive measurement using Vnew either through an inverse-matrix/matrix-vector computation with Dnew or through best-match row selection against D, with calibrated values derived from comparison with entries of the invasive column vector Y.
Stated Advantages
Allows measuring the bioparameter using a non-invasive bioparameter reading at a pre-defined level of acceptability.
Provides calibrated bioparameter values for the patient by using learning-matrix correlations between non-invasive bioparametric readings and invasive bioparametric readings.
Documented Applications
Monitoring a bioparameter of a patient using a bioparameter monitoring device having an invasive component and a non-invasive component.
A portable bioparameter-monitoring medical device usable by a patient for monitoring a bioparameter of the patient.
Producing a portable bioparameter-monitoring medical device custom-tailored to a patient by calibrating the non-invasive component using repeated invasively and non-invasively measured paired readings within proximity time.
Bioparameter monitoring software executed by digital processors to perform learning and calibration to output a calibrated bioparameter value.
Interested in licensing this patent?