SI2-SSE: SCIFIO: An Extensible Framework for Scientific Image Interoperability

University Of Wisconsin-Madison, Madison WI

Investigators

Abstract

Digital imaging is one of the most commonly used tools in all of science. In a wide array of disciplines ranging from astronomy and geology to environmental science and biology, digital imaging approaches are used to capture dynamic processes in great detail. In the biological sciences, digital light microscopy has transformed the field with unprecedented levels of accessibility, functionality and performance with fewer compromises. Not only have there been great advances in the hardware and software needed to collect digital images, but also in the software tools for analysis, interpretation, storage and dissemination of the images. The accessibility of digital microscopy and imaging in general has resulted not only in widespread use and adoption but in increased innovation as well, with scientist pushing the envelope in developing core technologies in new directions such as hybrid and multimodal developments. Despite the great technological advances occurring in the field by commercial and academic researchers alike, there still is a fundamental barrier in imaging research that nearly every scientist encounters in their imaging workflow: the Proprietary File Format (PFF). PFFs are what the majority of software programs use to record images and any subsequent image analysis. While there are implemented efforts to create better open microscopy formats to help reduce the number of PFFs, it is clear that one universal format for everyone is not practical. Rather, the greatest practical need facing the community is not a universal scientific imaging format, but rather a universal scientific imaging format converter. With such a system, any current or future imaging format can be supported, including conversion from any PFF to any open standard. A universal imaging converter would enable a scientist to open a PFF from any imaging system and fully parse and analyze the full image contents without the need for any proprietary software. Such a converter would not only be of great utility to biologists but also of great benefit to instrument developers, who are equally limited by the lack of transparency and access of PFFs. We propose to develop a robust scientific software element for imaging file format interoperability. This effort that we dub "SCIFIO" for "Scientific Image Format Input and Output" would build on our current successful "Bio-Formats" efforts to make a file converter for light microscopy in our research domain and "harden" these efforts to make a robust interchange library for all of biological microscopy. The system will be generalizable, extensible and adaptable to new emerging microscopy types. It will also serve as a model for adaptation to other scientific imaging types. A great practical barrier to collaborative work in imaging is the issue of proprietary file formats. One of the most fundamental needs in imaging is being able to open and freely share the original pixel information and associated text information with others using any processing software or workflow desired. We are developing "Scientific Image Format Input and Output" (SCIFIO) a robust software package that can read and convert any proprietary image file format. By harnessing the power of a reusable software project like SCIFIO, the community will be able to freely share all content collected on any imaging system both for visualization and quantitative analysis. SCIFIO will be developed as a robust software element, both as a library anyone can utilize and as a full software tool kit that any developer can easily use to add converter support to their application freely. This is important as our target---and thus our impact---is not only on the research scientist, but the wider community including researchers from other fields, academic software developers, commercial software developers, and educators.