基于U-Net++的舌图像分割与优化方法研究  

Research on tongue image segmentation and optimization method based on U-Net++

作　　者：廖哲皓 刘忆宁 LIAO Zhehao;LIU Yining(Guilin University of Electronic Technology,Guilin 541004,China;Wenzhou University of Technology,Wenzhou 325027,China)

机构地区：[1]桂林电子科技大学,广西桂林541004 [2]温州理工学院,浙江温州325027

出　　处：《无线互联科技》2025年第6期20-24,共5页Wireless Internet Science and Technology

摘　　要：文章研究了一种基于U-Net++模型的舌图像分割方法,采用一种加权交叉熵损失函数来提升图像分割效果。在实验阶段,使用开源的舌图像数据集对方法进行测试。为了评估模型的综合性能,采用交并比和Dice系数作为评价指标。结果表明,采用加权交叉熵损失函数的模型在分割精度上有显著提升,尤其是在交并比和Dice系数2个指标上优于传统方法。该结果验证了加权交叉熵损失函数在舌图像分割任务中的有效性。The article investigates a tongue image segmentation method based on the U-Net++model and utilizes a weighted cross-entropy loss function to improve segmentation performance.During the experimental phase,the method is tested using an open-source tongue image dataset.To evaluate the overall performance of the model,the Intersection Over Union(IoU)and Dice coefficient are used as evaluation metrics.The results indicate that the model using the weighted cross-entropy loss function shows significant improvement in segmentation accuracy,especially outperforming traditional methods in both IoU and Dice coefficient.This result confirms the effectiveness of the weighted cross-entropy loss function in the tongue image segmentation task.

关 键 词：U-Net++模型 舌图像 图像分割 损失函数 加权优化 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]