Abstract on the Open Microscopy Environment: Difference between revisions
m (Created page with 'Michael Calhoun and Ilya Goldberg: Image Language Processing and Encoding Abstract The Open Microscopy Environment (OME) has a highly specified data model (including acquisitio...') |
mNo edit summary |
||
| Line 3: | Line 3: | ||
Abstract | Abstract | ||
The Open Microscopy Environment (OME) has a highly specified data model (including acquisition, workflow and annotation), parts of which are in use by thousands of labs worldwide. We contrast lessons-learned from design and application of OME, including potentially reusable aspects of this model, with more recent multi-disciplinary developments based on machine-learning. Bottom-up design, starting with the actual image data, maximizes usability and helps uncover (and utilize) application commonalities. The pragmatic end-goal of population-level data-mining includes the need to incorporate contextual and derived results with minimal divergence, encoding the image language into external vocabularies. The use of automated tag creation via pattern recognition is growing and will increase utilization by reducing data entry burden. We will present a framework with successful application across cell biology, pathology and radiology as an example to inform this process, and discuss how technology advances can impact design and use. | The [https://www.openmicroscopy.org/site Open Microscopy Environment] (OME) has a highly specified data model (including acquisition, workflow and annotation), parts of which are in use by thousands of labs worldwide. We contrast lessons-learned from design and application of OME, including potentially reusable aspects of this model, with more recent multi-disciplinary developments based on machine-learning. Bottom-up design, starting with the actual image data, maximizes usability and helps uncover (and utilize) application commonalities. The pragmatic end-goal of population-level data-mining includes the need to incorporate contextual and derived results with minimal divergence, encoding the image language into external vocabularies. The use of automated tag creation via pattern recognition is growing and will increase utilization by reducing data entry burden. We will present a framework with successful application across cell biology, pathology and radiology as an example to inform this process, and discuss how technology advances can impact design and use. | ||
Latest revision as of 21:02, 2 June 2014
Michael Calhoun and Ilya Goldberg: Image Language Processing and Encoding
Abstract
The Open Microscopy Environment (OME) has a highly specified data model (including acquisition, workflow and annotation), parts of which are in use by thousands of labs worldwide. We contrast lessons-learned from design and application of OME, including potentially reusable aspects of this model, with more recent multi-disciplinary developments based on machine-learning. Bottom-up design, starting with the actual image data, maximizes usability and helps uncover (and utilize) application commonalities. The pragmatic end-goal of population-level data-mining includes the need to incorporate contextual and derived results with minimal divergence, encoding the image language into external vocabularies. The use of automated tag creation via pattern recognition is growing and will increase utilization by reducing data entry burden. We will present a framework with successful application across cell biology, pathology and radiology as an example to inform this process, and discuss how technology advances can impact design and use.
