Skip to main content

Chunkflow: hybrid cloud processing of large 3D images by convolutional nets

Author(s): Wu, Jingpeng; Silversmith, William M; Lee, Kisuk; Seung, H Sebastian

To refer to this page use:
Abstract: Automated microscopes with both high resolution and large field of view are generating terascale and even petascale 3D images. A local cluster might not have enough computational resources to process them in reasonable time, but public cloud platforms can provide computational resources on demand. Convolutional networks have become the state-of-the-art approach for 3D biological image analysis1,2 , and cloud processing by 3D convolutional nets has been used for processing independent small image stacks3–5 . However, cloud computing tools to perform distributed processing of terascale or petascale 3D images by convolutional nets are lacking. Here, we report chunkflow, a framework for distributing computational tasks over both cloud and local computational resources, including both GPUs and CPUs with multiple deep-learning framework back ends, to maximize efficiency, increase flexibility and reduce cost.
Publication Date: 2021
Citation: Wu, Jingpeng, William M. Silversmith, Kisuk Lee, and H. Sebastian Seung. "Chunkflow: hybrid cloud processing of large 3D images by convolutional nets." Nature Methods 18, no. 4 (2021): pp. 328-330. doi:10.1038/s41592-021-01088-5
DOI: 10.1038/s41592-021-01088-5
ISSN: 1548-7105
Pages: 328 - 330
Type of Material: Journal Article
Journal/Proceeding Title: Nature Methods
Version: Author's manuscript

Items in OAR@Princeton are protected by copyright, with all rights reserved, unless otherwise indicated.