Quality control of visual diagnostics using neural network technologies

Stack of technologies

Machine Learning + Computer Vision

stack: C#, UWP, Postgres, Accord.Net, ContourAlalysis, Python, SciKit


Quality control of visual diagnostics using neural network technologies.

Machine Learning + Computer Vision

status: research project, developed, tested, implemented a pilot project.

Tasks and problems

Actual risks when using the device  (human factor):

        Admissibility of error when using medical equipment.

        Data interpretation from images survey.

        30% of incorrect surveys as a result of not following the inspection procedure and making an error in the use of medical equipment.

*The analyzed data from fibroscans in 4 hepato-centers for the period September  2013 were analyzed. - July 2015.

Technological solution of the problem

By connecting our device to the ultrasound machine and machine learning algorithms, our program during the ultrasound confirms the results of the scan and helps to exclude the human factor in the diagnosis results.

Principle of operation