Skip to main content
. 2015 Aug 13;9:20. doi: 10.3389/fninf.2015.00020

Figure 3.

Figure 3

An illustration of the distributed processing paradigm. Raw image data (1) is divided into cuboids (2) based on user-specified parameters, with the necessary padding to perform computation. After processing (3), the annotations are inscribed and uploaded to OCP (4). Finally, processes are merged across block seams (5), using a similarity metric of the userś choice.