Method and apparatus for non-invasive glucose measurement
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-10687739-B2
Publication Date
2020-06-23
Expiration Date
Abstract
An apparatus for monitoring blood glucose comprising an invasive component for invasively measuring blood glucose and a non-invasive component, including color image sensor(s) to generate images from absorption of light that traversed the tissue, to receive a body part and generate a non-invasive blood glucose reading. Processor(s) convert the images into a vector V associated with a particular at least one invasive blood glucose measurement gk1, form a regular learning matrix, λ, implement a noninvasive isolation mechanism of the tissue glucose level by unique association of the vector Vk with an invasive blood glucose level, determine a neural network from the learning set λ by pairing vectors into a branch and forming multiple branches into loops, wherein two vectors are paired if they have a pre-defined similarity in the blood glucose levels that each are associated with; and calibrate the neural network by having it pass at least one test.
Core Innovation
The described invention monitors a person’s blood glucose by invasively measuring blood glucose using an invasive component, storing invasive blood glucose measurement gk1 in a non-invasive component, and generating a series of images reflecting absorption of light that traversed tissue of a body part. The series of images is converted into a vector V, where the vector V is associated with the particular at least one invasive blood glucose measurement gk1 describing a momentary glucose level. A set of vectors associated with invasively determined blood glucose levels defines a learning set λ of the device.
From the learning set λ, the device forms an M by N regular learning matrix λ by repeating invasive measurement and image generation so as to acquire N vectors, with N determined according to the person’s invasively determined blood glucose level. The invention implements a noninvasive isolation mechanism of a tissue glucose level by a unique association of a vector Vk with an invasive blood glucose level, defined as Vk→gk, including a condition that if Vk=Vn then gk=gn for any k≠n. Vectors Vk that fail this isolation association are discarded.
The invention determines a neural network from the learning set λ by pairing vectors into a branch and forming multiple branches into loops, where two vectors are paired if they have a pre-defined similarity in the blood glucose levels that each are associated with. The neural network is calibrated by passing at least one internal blind test, and the device is further refined by internal blind-test procedures. Additional refinement can include using vectors from newly measured body-location image series after calibration to associate a new vector Vnew via an optimal correlated loop of the calibrated neural network.
Claims Coverage
The independent claims cover a method, an apparatus, and a non-transitory computer-readable medium for non-invasive blood glucose monitoring based on four inventive features.
Invasive-to-image learning set and regular learning matrix formation
An invasively measuring blood glucose measurement gk1 and generating image series reflecting absorption of light through tissue, converting the series of images into a vector V associated with the invasive measurement to define a learning set λ, and forming from learning vectors an M by N regular learning matrix λ by repeating so as to acquire N vectors with N determined according to the person’s invasively determined blood glucose level.
Noninvasive isolation mechanism with unique vector-to-glucose association and discarding
Implementing a noninvasive isolation mechanism of a tissue glucose level by unique association of vector Vk with an invasive blood glucose level defined as Vk→gk, including a condition that if Vk=Vn then gk=gn for any k≠n, and discarding any vectors Vk that fails the noninvasive isolation mechanism’s association.
Neural network with branch-and-loop pairing by pre-defined glucose similarity
Determining a neural network from learning set λ by pairing vectors into a branch and forming multiple branches into loops, where two vectors are paired if they have a pre-defined similarity in the blood glucose levels that each are associated with.
Internal blind-test calibration of the neural network
Calibrating the neural network by having the neural network pass at least one internal blind test.
Across the independent claims, the core inventive coverage is the combination of invasive measurements paired with tissue light-absorption image-derived vectors to build a learning set and M×N regular learning matrix, enforcing a unique vector-to-invasive-glucose isolation association with discarding of failing vectors, constructing a neural network with branch-and-loop pairing based on pre-defined similarity of associated glucose levels, and calibrating via at least one internal blind test.
Stated Advantages
Documented Applications
No documented applications found
Interested in licensing this patent?