作者: Martin Welk , Christian Feddern , Bernhard Burgeth , Joachim Weickert
DOI: 10.1007/978-3-540-45243-0_3
关键词:
摘要: Novel matrix-valued imaging techniques such as diffusion tensor magnetic resonance require the development of edge-preserving nonlinear filters. In this paper we introduce a median filter for tensor-valued data. We show that it inherits number favourable properties from scalar-valued filtering, and present experiments on synthetic well real-world images illustrate its performance.