A fully automatic geometric parameters determining method for electron tomography

Abstract

Electron tomography (ET) is a promising technique for investigating in situ three-dimensional (3D) structure of proteins and protein complexes. To obtain a high-resolution 3D ET reconstruction, alignment and geometric parameters determination of ET tilt series are necessary. However, the common geometric parameters determining methods depend on human intervention, which are not only fairly subjective and easily introduce errors but also labor intensive for high-throughput tomographic reconstructions. To overcome these problems, in this paper, we presented a fully automatic geometric parameters determining method. Taking advantage of the high-contrast reprojections of ICON and a series of image processing and edge recognition techniques, our method achieves a high-precision full automation for geometric parameters determining. Experimental results on the resin embedded dataset show that our method has a high accuracy comparable to the common ‘manual positioning’ method.