Abstract
Due to the limited field of view of the microscopes, acquisitions of macroscopic specimens require many parallel image stacks to cover the whole volume of interest. Overlapping regions are introduced among stacks in order to make it possible automatic alignment by means of a 3D stitching tool. Since state-of-the-art microscopes coupled with chemical clearing procedures can generate 3D images whose size exceeds the Terabyte, parallelization is required to keep stitching time within acceptable limits. In the present paper we discuss how multi-level parallelization reduces the execution times of TeraStitcher, a tool designed to deal with very large images. Two algorithms performing dataset partition for efficient parallelization in a transparent way are presented together with experimental results proving the effectiveness of the approach that achieves a speedup close to 300×, when both coarse- and fine-grained parallelism are exploited. Multi-level parallelization of TeraStitcher led to a significant reduction of processing times with no changes in the user interface, and with no additional effort required for the maintenance of code.
Highlights
State-of-the-artmicroscopes (Dodt and et al, 2007; Silvestri et al, 2012), coupled with chemical clearing procedures to render brain tissue transparent (Chung et al, 2013) can generate 3D images having size in the order of Terabyte (TB) at high throughput
We have presented an approach aiming at reducing the stitching time of TeraStitcher using parallel processing at different grain levels
Additional advantages of the proposed approach are that parallelization does not require additional efforts for the maintenance of TeraStitcher code, no special image file formats are introduced, there are no changes in the user interface, apart the need to specify the desired level of parallelism, and execution on distributed memory platforms is possible
Summary
State-of-the-artmicroscopes (Dodt and et al, 2007; Silvestri et al, 2012), coupled with chemical clearing procedures to render brain tissue transparent (Chung et al, 2013) can generate 3D images having size in the order of Terabyte (TB) at high throughput. Processing and manipulation of these images require new software tools to perform a number of functions from stitching to visualization, to analysis. Due to the limited field of view of the microscopes, acquisitions of macroscopic specimens (e.g., whole mammalian brains) require many parallel image stacks ( referred to as tiles in the following) to cover the whole volume of interest. Multiple tiles, each composed by thousands of slices, are acquired using motorized stages. For volumes of approximately 1 cm at sub-micrometer resolution, the size of acquired data may exceed the Teravoxel. Since tile positions provided by the stages are not sufficient to determine a reliable displacement between tiles, an overlapping region is introduced with the purpose of making possible the automatic combination of the tiles by means of a stitching software tool
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
