SmartPoct – Intelligent Text Recognition for Medical POCT Devices

eHealth Technologies Improving Primary Care

SmartPoct – Intelligent Text Recognition for Medical POCT Devices

eHealth Technologies Improving Primary Care

SmartPoct – Intelligent Text Recognition for Medical POCT Devices

eHealth Technologies Improving Primary Care
The Client
POCT (point of care testing)
PyTorch, OpenCV, Numpy UNet, Xception
PyTorch, OpenCV, Numpy UNet, Xception
PyTorch, OpenCV, Numpy UNet, Xception

Challenge

POCT (point of care testing) is a portable medical device that is used to perform lab testing at a patient's bedside. They usually have closed communication protocols, which vary from one type of device to another.

The task was to create an image recognition system for:

Challenge

POCT (point of care testing) is a portable medical device that is used to perform lab testing at a patient's bedside. They usually have closed communication protocols, which vary from one type of device to another.

The task was to create an image recognition system for:

Challenge

POCT (point of care testing) is a portable medical device that is used to perform lab testing at a patient's bedside. They usually have closed communication protocols, which vary from one type of device to another.

The task was to create an image recognition system for:
independent recording of POCT test results in the system
sending lab results to the server without using medical protocols
independent recording of POCT test results in the system
sending lab results to the server without using medical protocols
independent recording of POCT test results in the system
sending lab results to the server without using medical protocols

Approach

Approach

Approach


After performing a test on a patient, the device displays a certain value. A nurse or lab assistant takes a photo of the screen. Our Machine Learning system is able to recognize the required value of the test result, read it and send it to the server so it becomes available in the patient's medical record. This text recognition technology works independently of POCT device model and the protocols used. Thus, laboratories receive POCT results easily and safely in to their lab-system (LIS).

After performing a test on a patient, the device displays a certain value. A nurse or lab assistant takes a photo of the screen. Our Machine Learning system is able to recognize the required value of the test result, read it and send it to the server so it becomes available in the patient's medical record. This text recognition technology works independently of POCT device model and the protocols used. Thus, laboratories receive POCT results easily and safely in to their lab-system (LIS).

After performing a test on a patient, the device displays a certain value. A nurse or lab assistant takes a photo of the screen. Our Machine Learning system is able to recognize the required value of the test result, read it and send it to the server so it becomes available in the patient's medical record. This text recognition technology works independently of POCT device model and the protocols used. Thus, laboratories receive POCT results easily and safely in to their lab-system (LIS).
SmartPOCT helps primary care providers:
Recognize test results
Enter them in medical records
Exclude human processing
Save time
Improve patient care
SmartPOCT helps primary care providers:
Recognize test results
Enter them in medical records
Exclude human processing
Save time
Improve patient care
SmartPOCT helps primary care providers:
Recognize test results
Enter them in medical records
Exclude human processing
Save time
Improve patient care

Solution

ENBISYS has built a pipeline of models based on deep learning:

Solution

ENBISYS has built a pipeline of models based on deep learning:

Solution

ENBISYS has built a pipeline of models based on deep learning:
image
segmentation
image
classification
affine
transformations
optical character recognition (OCR) with bi-directional LSTM
image
segmentation
image
classification
affine
transformations
optical character recognition (OCR) with bi-directional LSTM
image
segmentation
image
classification
affine
transformations
optical character recognition (OCR) with bi-directional LSTM
Outcome
The system currently works with 4 types of POCT devices. It helps automate certain procedures and exclude manual processing of data. It offers additional advantages for both patients and care providers and also helps laboratories operate faster.
Outcome
The system currently works with 4 types of POCT devices. It helps automate certain procedures and exclude manual processing of data. It offers additional advantages for both patients and care providers and also helps laboratories operate faster.
Outcome
The system currently works with 4 types of POCT devices. It helps automate certain procedures and exclude manual processing of data. It offers additional advantages for both patients and care providers and also helps laboratories operate faster.
Let's discuss your Case!
Let's discuss your Case!
Let's discuss your Case!