人工智能培训

搜索

人工智能论文:胶囊回溯及其在语义分割中的应用(Trace-back Along Capsules and Its Application on Semantic

[复制链接]
wynrefer 发表于 2019-1-11 11:04:19 | 显示全部楼层 |阅读模式
wynrefer 2019-1-11 11:04:19 124 0 显示全部楼层
人工智能论文:胶囊回溯及其在语义分割中的应用(Trace-back Along Capsules and Its Application on Semantic Segmentation)在本文中,我们提出了一种基于胶囊的神经网络模型来解决这些语义分割问题。通过利用胶囊层中可用的可萃取部分 - 整体依赖性,我们通过递归的逐层过程推导出各个胶囊的类标签的概率。我们将此过程建模为回溯管道,并将其作为构建端到端分段网络的核心部分。以明确的方式共同寻求所提出的框架,图像级类标签和对象边界,这对于现有技术的完全卷积网络(FCN)解决方案具有显着的优势。在改进的MNIST和神经图像上进行的实验表明,与领先的FCNvariant相比,我们的模型显着提高了分割性能。
In this paper, we propose a capsule-based neural network model to solve thesemantic segmentation problem.By taking advantage of the extractablepart-whole dependencies available in capsule layers, we derive theprobabilities of the class labels for individual capsules through a recursive,layer-by-layer procedure.We model this procedure as a traceback pipeline andtake it as a central piece to build an end-to-end segmentation network.Underthe proposed framework, image-level class labels and object boundaries arejointly sought in an explicit manner, which poses a significant advantage overthe state-of-the-art fully convolutional network (FCN) solutions.Experimentsconducted on modified MNIST and neuroimages demonstrate that our modelconsiderably enhance the segmentation performance compared to the leading FCNvariant.人工智能论文:胶囊回溯及其在语义分割中的应用(Trace-back Along Capsules and Its Application on Semantic Segmentation) s2aRaOO6sZ6Br9A9.jpg
URL地址:https://arxiv.org/abs/1901.02920     ----pdf下载地址:https://arxiv.org/pdf/1901.02920    ----人工智能论文:胶囊回溯及其在语义分割中的应用(Trace-back Along Capsules and Its Application on Semantic Segmentation)
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则 返回列表 发新帖

wynrefer当前离线
新手上路

查看:124 | 回复:0

快速回复 返回顶部 返回列表