基于改进U-Net神经网络的人体血细胞计数

Human Blood Cell Counting Based on Improved U-Net Neural Network

自动血细胞计数在医学领域具有重要意义,现有传统方法在计数过程中存在一定误差。针对传统U-Net模型进行优化分别提出了Res-U-Net模型和VGG-U-Net模型,二者均提高了人体血细胞计数的精度。首先实验采集的人体血细胞图像制作成数据集,然后将优化模型与传统U-Net模型对比分析,实验结果表明在传统U-Net细胞计数精度为92%的基础上,Res-U-Net提升到94%,VGG-U-Net提升到95%,二者均显著提高了计数精度。实验数据很好验证了两种优化模型的有效性,为人体血细胞计数领域的自动化技术提供了新的思路。

Automatic blood cell counting is of great significance in the medical field,and existing traditional methods have certain errors in the counting process.This article proposes Res-U-Net model and VGG-U-Net model for optimizing traditional U-Net models,both of which improve the accuracy of human blood cell counting.Firstly,the human blood cell images collected in the experiment are made into a dataset,and then the optimized model is compared and analyzed with the traditional U-Net model.The experimental results show that on the basis of the traditional U-Net cell counting accuracy of 92%,Res-U-Net is improved to 94%and VGG-U-Net is improved to 95%,both of which significantly improve the counting accuracy.The experimental data has effectively verified the effectiveness of the two optimization models,providing new ideas for automation technology in the field of human blood cell counting.