基于综合的全变差贝叶斯学习
Total Variation Bayesian Learning via Synthesis
摘要:
针对信号与图像处理中的逆问题,提出了一种基于总变分(TV)稀疏先验的稀疏贝叶斯学习算法。由于使用的先验,以及先验参数直接从数据中估计的事实,稀疏贝叶斯学习通常比用于稀疏信号恢复的典型最大后验贝叶斯估计产生更准确的结果。它还提供了一个完整的后验分布。然而,稀疏贝叶斯学习仅适用于具有直接稀疏先验或通过合成形成的问题。本文演示了如何与电视稀疏先验的问题,可以制定在一个综合的方法。然后,我们开发了一种方法,将这种基于合成的电视与稀疏贝叶斯学习算法相结合,并提供数值例子来演示我们的新技术是如何有效地采用。
英文摘要:
This paper presents a sparse Bayesian learning algorithm for inverse problems in signal and image processing with a total variation (TV) sparsity prior. Because of the prior used, and the fact that the prior parameters are estimated directly from the data, sparse Bayesian learning often produces more accurate results than the typical maximum a posteriori Bayesian estimates for sparse signal recovery. It also provides a full posterior distribution. However, sparse Bayesian learning is only available to problems with a direct sparsity prior or those formed via synthesis. This paper demonstrates how a problem with a TV sparsity prior can be formulated in a synthesis approach. We then develop a method that combines this synthesis-based TV with the sparse Bayesian learning algorithm and provide numerical examples to demonstrate how our new technique is effectively employed.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。