Multi-task learning cell counting method based on convolutional neural network

Abstract

The invention discloses a multi-task learning cell counting method based on a convolutional neural network, which is suitable for carrying out cell counting in a biological cell microscopic image withrelatively dense cells and relatively large quantity. The method comprises the following steps: preprocessing the biological cell image data set to obtain a training set and a test set; constructinga convolutional neural network model of cell counting based on multi-column feature map fusion and multi-task learning; training the convolutional neural network model, using the training set after pretreatment and the network model, and through the propagation algorithm and parameter updating, obtaining the optimized model weight parameters; testing the convolutional neural network model, using the pretreated test set and the weight parameters of the optimal network model, testing the cell picture, getting the output cell density map and the number of cell estimates, and performing evaluation. The method can improve the performance of biological cell counting and improve the accuracy rate.

Description

technical field [0001] The invention relates to the technical fields of computer vision and deep learning, in particular to a convolutional neural network-based multi-task learning cell counting method. Background technique [0002] In current medical and biological research, the study of biological cells has become a key technology, which is an important tool for studying pathogenesis and biological processes. The current trend is that more and more researchers have higher goals for the research and analysis of microscopic images of biological cells. In order to achieve their research goals, they need to process and analyze cells through various computer image technologies. Processing includes cell detection, segmentation, counting, and more. Among them, cell counting technology has been more and more widely used. In medicine, many diseases and drug research need to know the number of certain specific cells: on the one hand, it can be based on the number of target cells in...

AI Technical Summary

Problems solved by technology

This detection-based counting method can have high accuracy after training, but it is limited to images with rich cell features and a small number of cells, and the efficiency of detecting one by one is low, so the researchers established for the cell counting task The purpose of the regression model is to learn the direct mapping between the features in the image and the number of cells. In the training phase, the cell map and annotation information in the training set are used to obtain the mapping relationship between the cells and the number. In the testing phase, the cell number is directly obtained according to the input image. quantity estimate

Images