福建电脑2024，Vol.40Issue(5)：1-5,5.DOI:10.16707/j.cnki.fjpc.2024.05.001
融合残差反卷积的图像分割算法研究
Research on Image Segmentation Algorithm based on Residual Deconvolution
摘要
Abstract
This paper proposes an image segmentation algorithm RDM-FCN that integrates residual deconvolution to address the issue of misclassification in FCN algorithm for processing complex scenes.In the encoder section,the VGG16 network is used to extract image features;In the decoder section,residual deconvolution modules are constructed and residual connections are introduced to enhance the transmission of cross layer features.By using the cross entropy loss function,the segmentation accuracy of the model is improved.The test results show that compared with the FCN algorithm,the accuracy of our algorithm has improved by 0.0347,the average intersection to union ratio has increased by 0.0215,and the average pixel accuracy has increased by 0.005.The experimental results show that the segmentation accuracy of the algorithm proposed in this paper is high,and it can effectively preserve the information of object edges and details.关键词
FCN网络/图像分割/残差反卷积/算法Key words
FCN Network/Image Segmentation/Residual Deconvolution Model/Algorithm分类
信息技术与安全科学引用本文复制引用
何松,唐程华,陈鑫..融合残差反卷积的图像分割算法研究[J].福建电脑,2024,40(5):1-5,5.基金项目
本文得到江西省研究生创新专项(No.YC2023-S662)资助. （No.YC2023-S662）
