A Picture Worth Six Words… and $3.4 Million

A research team including CyVerse leadership has been awarded $3.4 Million to develop cyberinfrastructure for large-scale image processing.

By Shelley Littin

Large-Scale IMage Processing Infrastructure Development, or LIMPID, is the six-word name of the National Science Foundation’s Office of Advanced Cyberinfrastructure new $3.4 Million grant to relieve a bottleneck in scientific image processing.

Nirav Merchant, CyVerse co-principal investigator and a member of the BIO5 Institute at the University of Arizona (UA), is a co-principal investigator of the project.

LIMPID, led by principal investigator B.S. Manjunath at the University of California-Santa Barbara (UCSB), aims “to create a large scale distributed image processing infrastructure, the LIMPID, through a broad, interdisciplinary collaboration of researchers in databases, image analysis, and sciences,” according to the grant announcement.

“The UCSB and CyVerse teams have been collaborating in addressing large-scale image informatics challenges,” said Manjunath, Director at the Center for Multimodel Big Data Science and Healthcare at UCSB. “The BisQue image informatics platform developed at UCSB is offered as a core service at CyVerse, and now the new LIMPID project will further strengthen these collaborative efforts between the two groups while diversifying into materials and marine sciences research.”

LIMPID’s image informatics abilities will be built upon the underlying computational infrastructure developed by CyVerse at the UA’s BIO5 Institute.

“Image data is a key component of many modern, large-scale scientific datasets,” said Merchant. “At CyVerse, we aim to facilitate all forms of data management. This award provides an excellent avenue to leverage the national cyberinfrastructure concepts we have developed at CyVerse to bring distributed image processing to researchers nationwide.”

The main goal of LIMPID, Merchant continued, “is to bring specific user communities – materials science, marine science, and neuroscience – the ability to share, test, and refine methods that have common underlying algorithms and procedures.”

The grant investigators aim to streamline the processes of image analysis, processing, and annotation for distributed research teams. Through LIMPID, the infrastructure and interface for research questions that would otherwise require expensive software will become widely available to researchers across the country.

“For the broader scientific community, this immediately provides a resource for joint data and methods publication,” the grant authors write. “The project will prepare a diverse cadre of students and researchers, including women and members of underrepresented groups, to tackle complex problems in an interdisciplinary environment.”

“We will focus on making contemporary machine-learning methods easier to use for domain scientists,” said Merchant, who is also Director of BioComputing at Arizona Research Laboratories, “allowing them to build models and training datasets from their curated images, and use those training models to find patterns and objects.”

The accessibility may help to engage minorities and underrepresented groups in science, which might not otherwise have access to advanced image processing capabilities.

“LIMPID will further the mission that CyVerse began,” Merchant added, “of making cyberinfrastructure broadly accessible to enable researchers to share data, easily communicate, and complete their large-scale data projects, making their results more immediately applicable to the broader scientific communities.”