Non-invasive blood glucose monitoring system

Diabetes mellitus is a severe disease characterized by elevated blood glucose levels due to dysregulation of the hormone insulin. It requires careful monitoring involving constant measurement of the hormone.

Based on a lot of trial and error, a team from Kennesaw State University (USA) created a non-invasive process that can identify the value of blood glucose without the need to take a sample. The process uses light that passes through human tissue in the ear or finger and a small camera to capture images on the other side. Scientists then use a model to study the amount of light absorption in those images to determine the glucose concentration.

The system is based on Raspberry Pi, a portable camera (Raspberry Pi), and a visible light laser. The camera captures images when a visible-light laser passes through skin tissue. Glucose concentration is estimated by an artificial neural network model using light absorption and scattering. The prototype was developed using TensorFlow, Keras, and Python.

The accuracy of the prototype is 79%. If images are taken of the ear, it dims to 62%. Although the current data set is limited, the results are encouraging. However, in future studies, it is necessary to address three main limitations: increasing the size of the database to improve the robustness of the artificial neural network model, analyzing the impact of external factors such as skin color, skin thickness, and room temperature in the current prototype and improve the object to be suitable in the easy placement of fingers and ears.

This apparatus demonstrates that blood glucose concentration can be estimated using hardware that uses infrared images of human tissue. The team has tested the process on almost 50 people, but before filing a complete patent, they will test how the process works on people with different pigmentations and skin thicknesses.

The system is based on Raspberry Pi, a portable camera (Raspberry Pi camera), and a visible light laser. The camera captures images when a visible-light laser passes through skin tissue. Glucose concentration is estimated by an artificial neural network model using light absorption and scattering. The prototype was developed using TensorFlow, Keras, and Python.

The accuracy of the prototype is 79%. If images are taken of the ear, it dims to 62%. Although the current data set is limited, the results are encouraging. However, in future studies, it is necessary to address three main limitations: increasing the size of the database to improve the robustness of the artificial neural network model, analyzing the impact of external factors such as skin color, skin thickness, and room temperature in the current prototype and improve the object to be suitable in the easy placement of fingers and ears.

The system is based on Raspberry Pi, a portable camera (Raspberry Pi camera), and a visible light laser. The camera captures images when a visible-light laser passes through skin tissue. Glucose concentration is estimated by an artificial neural network model using light absorption and scattering. The prototype was developed using TensorFlow, Keras, and Python.

The accuracy of the prototype is 79%. If images are taken of the ear, it dims to 62%. Although the current data set is limited, the results are encouraging. However, in future studies, it is necessary to address three main limitations: increasing the size of the database to improve the robustness of the artificial neural network model, analyzing the impact of external factors such as skin color, skin thickness, and room temperature in the current prototype and improve the object to be suitable in the easy placement of fingers and ears.

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *