Morphological feature fused microscopic image bone marrow cell counting method and system

Abstract

The invention discloses a morphological feature fused microscopic image bone marrow cell counting method and system, and the method comprises the steps: 1) obtaining labeled and unlabeled bone marrow cell microscopic images, and carrying out the image preprocessing; 2) segmenting the labeled / label-free bone marrow cell microscopic image into a plurality of labeled / label-free single bone marrow cell microscopic images according to the cell segmentation marks and a watershed algorithm, and carrying out image preprocessing on the single bone marrow cell microscopic images; 3) carrying out cell nucleus and cytoplasm segmentation on the labeled and unlabeled single bone marrow cell microscopic images by using an improved maximum between-cluster variance method and an N-distance edge expansion method, and extracting morphological characteristics; 4) constructing a hybrid convolutional neural network model, and inputting the labeled single bone marrow cell microscopic image into the model for training; and 5) predicting cell categories of the unlabeled single bone marrow cell microscopic image by using the trained hybrid convolutional neural network model, and counting each cell category. The method has relatively high precision on bone marrow cell counting.

Description

technical field [0001] The invention relates to the technical field of counting bone marrow cells, in particular to a method and system for counting bone marrow cells using microscopic images that integrate morphological features. Background technique [0002] At present, there is an urgent need for accurate identification and counting of bone marrow cells in the field of bone marrow cell counting. The traditional identification and counting of bone marrow cells mainly starts from the morphological point of view of bone marrow cells, such as comprehensively considering the shape features, color features and texture features of bone marrow cells. Wait, give a rough judgment. In this case, the diagnostic results are highly dependent on the physician's level, and the diagnostic efficiency is low; the auxiliary diagnostic tool based on the extraction of bone marrow cell morphological features can improve the accuracy of cell identification and counting to a certain extent. The ...

AI Technical Summary

Problems solved by technology

The development of artificial intelligence technology has promoted the field of bone marrow cell counting to continue to mature, but how to combine traditional morphological feature methods and deep learning technology to improve the accuracy of bone marrow cell counting is still a challenge

Images