US 2020/0367833 A1 · PERSONALIZED NON-INVASIVE BLOOD GLUCOSE MEASUREMENT DEVICE USING MACHINE LEARNING OR DEEP LEARNING AND METHOD USING THE MEASUREMENT DEVICE

Abstract

A personalized non-invasive blood glucose measurement method according to the present disclosure includes: (a) collecting blood glucose value data of a patient by a processor unit; (b) determining an acceptable final error E′ with the blood glucose value data collected by the processor unit; (c) non-invasively predicting a blood glucose value in accordance with a setting mode by a patient when the final error E′ is determined, to measure a blood glucose value, by the processor unit; (d) periodically determining whether the final error E′ is calibrated by the processor unit; and (e) when it is not necessary to recalibrate in step (d), non-invasively predicting the blood glucose value in the range of the final error E′ determined in step (b) to measure the blood glucose value, to non-invasively provide an accurate personalized blood glucose value to a patient who needs to invasively measure a blood glucose every day.

Claims9

Dependent claims

A personalized non-invasive blood glucose measurement method, comprising:

(a) collecting blood glucose value data of a patient by a processor unit both invasively and non-invasively;
(b) determining an acceptable final error E′ with the non-invasive blood glucose value data collected by the processor unit;
(c) non-invasively predicting a blood glucose value in accordance with a setting mode by a patient when the final error E′ is determined, to measure a blood glucose value accurately, by the processor unit;
(d) periodically determining whether the final error E′ is calibrated by the processor unit;
and
(e) when it is not necessary to recalibrate in step (d), non-invasively predicting the blood glucose value in a range of the final error E′ determined in step (b) to measure the blood glucose value, by the processor unit.

3 more elements

A personalized non-invasive blood glucose measurement device using machine learning or deep learning, the device comprising:

a non-invasive blood glucose measurement unit which measures a blood glucose by a non-invasive sensor;
an invasive blood glucose measurement unit which measures a blood glucose by an invasive sensor;
a processor unit which collects an invasive blood glucose value and a non-invasive blood glucose value measured by the invasive blood glucose measurement unit and the non-invasive blood glucose measurement unit to calculate an error and generates a personalized error function model through the learning using machine learning or deep learning to compensate an error to calculate an expected non-invasive blood glucose value;
a mode selecting unit which selects a mode to output any one of the invasive blood glucose value, the non-invasive glucose value, and the expected non-invasive blood glucose value;
and
a display unit which displays any one of a selection mode by the mode selecting unit, patient information, time and date information, the invasive blood glucose value, the non-invasive blood glucose value, and the expected non-invasive blood glucose value.

3 more elements

Dependencies are inferred from the claim text.

Personalized non-invasive blood glucose measurement device using machine learning or deep learning and method using the measurement device

Life of this patent

Personalized non-invasive blood glucose measurement device using machine learning or deep learning and method using the measurement device

Life of this patent

Abstract

Claims9