%0 Journal Article
%T 基于U-Net++神经网络模型的细胞计数方法研究<br>Cell Counting Method Based on U-Net++ Neural Network Model
%A 刘钊
%A 代琦
%J Software Engineering and Applications
%P 163-173
%@ 2325-2278
%D 2023
%I Hans Publishing
%R 10.12677/SEA.2023.121017
%X 细胞计数是医学图像研究中一个重要的科学问题，精准的细胞计数能够检测潜在疾病与相关病变。针对现有计数方法中准确率不高和效率低的问题，本文提出了基于U-Net++神经网络模型的细胞计数方法。首先对图片进行增强与锐化处理，进而分割图片和提取其高维特征，构建基于U-Net++神经网络模型进行训练与分类，最终实现细胞计数。实验结果表明，本文提出的计数方法准确率可达到97.4%，改进后的U-Net++神经网络模型提供了一种端到端的细胞计数方法，通用性较强，更加符合实际需求。<br />
Cell count is an important scientific problem in medical image research. Accurate cell count can detect potential diseases and related lesions. Aiming at the problems of low accuracy and low efficiency in existing counting methods, this paper proposes a cell counting method based on U-net++ neural network model. Firstly, the image is enhanced and sharpened, then the image is segmtioned and its high-dimensional features are extracted, and the U-net++ neural network model is constructed for training and classification, and finally the cell count is realized. Experimental results show that the accuracy of the proposed counting method can reach 97.4%, and the improved U-net++ neural network model provides an end-to-end cell counting method, which has strong universality and is more in line with the actual needs.
%K 细胞计数，神经网络，U-Net++<br>Cell Counting
%K Neural Network
%K U-Net++
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=61968