Big imaging data batch processing
Dmitry V. Tuktarov 1, Egor V. Ilyukov 1, Dmitry A. Myagkov 1, Ivan V. Fedosov1, Oxana V. Semyachkina-Glushkovskaya2,3.
Affiliations:
1.Institute of Physics, , Saratov State University, Universitetskaya, 40, , 410012 Saratov, Russia.
2 Department of Biology, Saratov State University, Astrakhanskaya 82, 410012 Saratov, Russia.
3 Institute of Physics, Humboldt University, Newtonstrasse 15, 12489 Berlin, Germany.
Abstract
Fiji is the most popular open source software package widely used for image processing int the field of biomedical microscopy. However there is a memory management problem that causes an error with large batch processes. The issue is critical e.g. when long series of laser scanning microscope images are processed. Our goal was to optimize the Fiji macro structure to enable a reliable preview montage of several tens of gigabytes imaging data on PC. We successfully fixed the long standing memory issue with appropriate close of an image processing session and memory release procedure using basic Fiji macro commands. As a result our macro enables preview montage of 13.5 GB set of 12 images and requires only 3 GB of RAM while straightforward batch process typically results in memory out error at third image. We developed big image data preview macro that is not limited by number of images in dataset requires for batch process the same memory resources as that needed for one single image processing procedure.
This study was supported by grant from SRF 23-75-30001
Speaker
Dmitry V. Tuktarov
Institute of Physics, Saratov State University
Russia
Discussion
Ask question
