Image Processing Portal (IPP): A web-based solution for efficient processing of large-scale microscopy data
Nishanthi Dasanayaka1, Mark Endrei1, David Abramson1, James Springfield2, Nicholas Condon2, Deborah Barkauskas2 1Research Computing Centre, The University Of Queensland Australia2Institute for Molecular Bioscience, The University Of Queensland Australia
Abstract
Modern light-sheet microscopy techniques have made it possible to capture images of biological systems with high spatial resolution over lengthy time-lapse intervals, frequently resulting in the generation of terabytes of data in a single experiment. This drives the need for effective tools and expertise for image data management, processing, and analysis.
To address this challenge, we have developed Image Processing Portal (IPP), a web-based application in collaboration with the Metropolitan Data Caching Infrastructure (MeDICI) at the University of Queensland (UQ). IPP makes it possible for non-computer science researchers to perform microscopy image analysis using organizational High-Performance Computing (HPC) resources without the need for command-line expertise.
IPP streamlines tasks involving massively parallel image processing on GPU-accelerated HPC systems. The portal allows users to handle their experimental data on HPC clustered file systems interactively from within the browser. Users can process image files using a range of tools, including file converters, batch processing, and deskewing with a few mouse clicks. Users can download and execute macros from shared repositories to process image transformations such as Z-projection without needing to use the command line interface. Additionally, a custom metadata scraper is being implemented to accurately read additional image metadata to perform deconvolution all through a web interface.
With the IPP, any researcher with an interest in microscopy is now able to use high performance computing resources to perform image-processing tasks with terabyte-scale datasets without any of the experience required to use the command line tools.
Biography
Nishanthi Dasanayaka completed her PhD in Computer Science from Queensland University of Technology. She has specialized in system architecture design and algorithm development for time-sensitive road safety applications. She is currently working at the Research Computing Center, the University of Queensland. Her professional interests include incorporating cutting-edge web technology into the development of web apps to assist UQ researchers.