AI Deep Learning for Biometric Analyte Concentration Measurement

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional analyte concentration measurement methods face challenges in accurately distinguishing features and calibrating for interfering substances, particularly when multiple species other than hematocrit affect bioanalyte concentration, and struggle with rapid environmental changes, leading to inaccurate and delayed measurement results.

Innovation Solution

The method employs artificial intelligence deep learning using artificial neural networks to automatically extract features and calibrate formulas, classify analytes, and adapt to changing environments without the need for pre-defined features or extensive expert input, leveraging A-stepladder-type perturbation potentials and electrochemical biosensors to minimize interference and provide immediate accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional mathematical methods such as Multiple Linear Regression are used to estimate and reflect interference, then the measurement process is relatively simple and fast, but the accuracy deteriorates when multiple interfering species are combined or when unexpected interference patterns occur

Engineering Contradiction:
Improvemeasurement accuracyVSAvoidalgorithm complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent transforms the measurement approach by changing from conventional mathematical parameters (Multiple Linear Regression) to AI-based parameters (Deep Learning neural networks with multiple layers). This parameter change enables the system to handle complex interference patterns and multiple interfering species simultaneously, significantly improving measurement accuracy while maintaining computational efficiency through automated feature extraction and calibration.

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If pre-defined features and calibration curves are used, then the measurement process is straightforward, but the system cannot adapt to new interfering substances or environmental changes

Engineering Contradiction:
Improveadaptability to interfering substancesVSAvoidcalibration time
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent implements self-service through automated feature extraction and calibration using Deep Learning algorithms. The system automatically identifies relevant features from raw sensor signals and performs calibration without requiring manual intervention or pre-defined parameters. This enables the system to adapt to new interfering substances and environmental conditions automatically, eliminating the time loss associated with manual recalibration while maintaining high adaptability.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent applies preliminary action by performing extensive feature extraction, calibration, and interference pattern recognition during the training phase of the Deep Learning model. Once trained, the model is prepared in advance to handle various interfering substances and environmental conditions, enabling rapid adaptation during actual measurements without requiring time-consuming recalibration procedures.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If temperature equilibrium waiting time is required for sudden environmental changes, then measurement accuracy is maintained, but productivity decreases due to delayed measurements

Engineering Contradiction:
Improvetemperature stabilityVSAvoidmeasurement speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent replaces the mechanical/physical approach of waiting for thermal equilibrium with an AI-based computational approach. The Deep Learning model processes temperature variations and their effects on measurements in real-time, compensating for thermal effects through learned patterns rather than requiring physical equilibrium. This substitution enables immediate measurements even during sudden environmental changes while maintaining accuracy, significantly improving productivity without sacrificing measurement precision.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach enhances measurement accuracy and precision by automatically correcting for interfering substances and environmental changes, reducing calculation time to under 8 seconds and maintaining consistency across varying conditions, outperforming conventional methods in robustness and speed.

Implementation Method 1

an output electrical signal may be influenced due to a change in a diffusion coefficient toward an electrode or a reaction rate at an electrode surface due to a change in blood properties such as presence of a material other than an analyte oxidized at the electrode surface

Methodology Applied
Scientific EffectElectrochemical reaction: Oxidation

Implementation Method 2

obtaining a second sensitive current at two or more points of time by applying a A-stepladder-type perturbation potential after applying the constant DC voltage

Methodology Applied
Scientific EffectElectrical potential application: Electric Field

Data Source

PatentEP3822625B1Method for measuring concentration of a biometric measurement object by using artificial intelligence deep learning
Publication Date: 2023.09.27 I SENS INC
  • EP3822625B1 patent drawingFigure 1
  • EP3822625B1 patent drawingFigure 2
  • EP3822625B1 patent drawingFigure 3

AI summary

The present invention has been made in an effort to provide an analyte concentration measurement method using artificial intelligence deep learning, capable of extracting useful features that are not known in advance by humans through deep learning using artificial neural networks by imaging input signals obtained during the measurement time from samples with information (labels) to construct a data set and estimating a result value by applying an algorithm obtained through learning in this way, compared with conventional measurement techniques that are applied by devising formulas or methods that directly extract features for a long time by experts in related fields in order to extract effective features.