Three-dimensional feature-preserving noise reduction for real-time electron tomography

Electron tomography (ET) is the leading imaging technique for visualizing the molecular architecture of complex biological specimens. Currently, real-time ET systems allow scientists to acquire experimental datasets with the electron microscope and obtain a preliminary version of the three-dimensional structure of the specimen. In principle, this rough structure allows assessment of the quality of the sample and can also be used as a guide to collect more datasets. However, in practice, the low signal-to-noise ratio of the ET datasets precludes detailed interpretation and makes their assessment difficult. Therefore, noise reduction methods must be integrated in these real-time ET systems for their fully exploitation. This work proposes and evaluates two different multithreaded implementations of a sophisticated noise reduction method with capabilities of preservation of biologically relevant features. The exploitation of the computing power of modern multicore platforms makes this noise reduction method provide datasets appropriate for assessment in a matter of a few minutes, thereby making it suitable for integration in current real-time electron tomography systems.