Blood cell image detection and counting method based on convolutional neural network

Abstract

The invention relates to a blood cell image detection and counting method based on a convolutional neural network, and belongs to the field of medical image processing, and the method comprises the steps: dividing an obtained blood cell image data set into a training set, a verification set and a test set, and carrying out the enhancement of the blood cell image data set through employing a Mosaic algorithm; inputting the image into a CSPDarkNet53 network to obtain a plurality of feature maps, and transmitting the feature maps into a Neck network to extract fusion features; predicting the blood cell image by using an improved YOLOv4 object detection algorithm; performing confidence score sorting on the prediction frame, and obtaining a finally displayed prediction frame through a non-maximum suppression algorithm; then counting the blood cells by using the predictive tags of the cells; the prediction result of the detection model being verified again by adopting KNN and DIOU, and the problem of platelet repeated detection being eliminated. The method realizes accurate and rapid detection and counting of blood cells, and has great practical application value.

Description

technical field [0001] The invention belongs to the field of medical image processing, and relates to a blood cell image detection and counting method based on a convolutional neural network. Background technique [0002] Blood counts are an important test that laboratory physicians often perform to assess the health of patients. Blood cells mainly contain three types of cells, namely white blood cells, red blood cells and platelets. Among them, the most common cells in the blood are red blood cells, accounting for 40%-45% of the total number of blood cells; white blood cells, also known as white blood cells, only account for 1% of the total number of blood cells; The ratio is also great. The main function of red blood cells is to transport oxygen to various tissues of the body, so the number of red blood cells will affect the amount of oxygen received by various tissues. Blood cells are immune cells that fight various pathological infections. Platelets help blood clot. ...

AI Technical Summary

Problems solved by technology

In human blood, the number of these blood cells is very large. Traditional artificial blood cell counting is done by a hemocytometer. This counting method is not only very time-consuming, but also has a large error

At present, there is no research on the recognition and counting of blood cells in blood cell images using deep learning-based methods

Images