标题： 带有森林稀疏性的学习 显示英文标题

标题： Learning with Forest Sparsity

摘要： 在本文中，我们研究了一个称为\emph{森林稀疏性}的新模型，用于稀疏学习和压缩感知。它是当数据的支持集由一系列相互关联的树组成时的标准稀疏性的扩展。森林稀疏性存在于许多实际应用中，如多对比度 MRI、并行 MRI、多光谱图像和彩色图像恢复。我们从理论上证明了森林稀疏性的优势，即在压缩感知中成功恢复所需测量值要少得多。此外，提出了一种新算法，并应用于具有森林稀疏性的几个应用中。所有实验结果都验证了森林稀疏性的优越性。 显示英文摘要

摘要： In this paper, we investigate a new model called \emph{forest sparsity} for sparse learning and compressive sensing. It is an extension of standard sparsity when the support set of the data is consisted of a series of mutually correlated trees. Forest sparsity exists in many practical applications such as multi-contrast MRI, parallel MRI, multispectral image and color image recovery. We theoretically prove the benefit of forest sparsity, that much less measurements are required for successful recovery in compressive sensing. Moreover, a new algorithm is proposed and applied on several applications with forest sparsity. All experimental results validate the superiority of forest sparsity.