An image recognition application generally consists of six major steps:
- Importing images for analysis
- Enhancing the images to prepare for segmentation or measurement
- Segmenting the images to detect biological objects and/or define regions of interest
- Analyzing the data and classifying objects into subpopulations
- Exporting the results for archival or analysis with 3rd party software.
SVCell recipes encode the processing rules for the recognition application, and can be applied to automatically enhance, detect, segment, measure, classify and analyze cellular and subcellular phenotypes. Recipes enable SVCell to be general purpose (because they can be taught for any application), and yet also provide “one-click” execution and automation for repeated application on multiple image sets.
SVCell currently provides import, enhancement, segmentation, measurement and decision recipes. See the Using SVCell section for more information about what can be accomplished using recipes.
Individual recipes can be grouped into recipe lists and locked into any processing order to support a wide range of applications, including applications yet to be thought of! The recipe list can be deployed throughout the lab or organization. Recipes can be saved with their teaching images (see Teaching SVCell section), or without to reduce the file size or maintain confidentiality.