In order to automate microscopic examinations of female urogenital samples, the “learning” of the “Vision Cyto STD” system was performed using the technology of image recognition based on convolutional neural networks and deep learning.
More than a thousand digital slides with urogenital material with different microscopic composition were evaluated during the learning process. As the dataset was replenished with new clinical examples, the accuracy of object recognition increased.
The use of artificial neural networks (ANN) for imaging and analysis of urogenital samples ensures high quality of examination performance due to the clarity of identification, accuracy of analysis and correctness of entering the examination results. Using of ANN speeds up the process of giving the results, allows to minimize the influence of the subjective factor, and also provides an opportunity to redistribute the workload of laboratory workers, without expanding the laboratory staff when the number of examinations to be performed increases.
A brief overview of the automated Vision Cyto STDs system is presented.