Electron tomography (ET) has recently afforded new insights into neuronal architecture. However, the tedious process of sample preparation, image acquisition, alignment, back projection, and additional segmentation process of ET repels beginners. We have tried Hitachi's commercial packages integrated with a Hitachi H-7650 TEM to examine the potential of using an automated fiducial-less approach for our own neuroscience research. Semi-thick sections (200–300 nm) were cut from blocks of fixed mouse (C57BL) cerebellum and prepared for ET. Sets of images were collected automatically as each section was tilted by 2° increments (±60°). “Virtual” image volumes were computationally reconstructed in three dimension (3D) with the EMIP software using either the commonly used “weighted back-projection” (WBP) method or “topography-based reconstruction” (TBR) algorithm for comparison. Computed tomograms using the TBR were more precisely reconstructed compared with the WBP method. Following reconstruction, the image volumes were imported into the 3D editing software A-View and segmented according to synaptic organization. The detailed synaptic components were revealed by very thin virtual image slices; 3D models of synapse structure could be constructed efficiently. Overall, this simplified system provided us with a graspable tool for pursuing ET studies in neuroscience.