Data+ is a project conducted by select faculty and students at Duke University. The project is designed to produce an automated process for anatomical annotation of lung tissue images. Image data and descriptive metadata used in the project are from research centers working in the LungMAP consortium. To develop algorithms for machine based annotation, the Duke students worked with staff conducting a similar effort at LungMAP Research Center, CCHMC. Data+ generated de novo segmentation rules based on metadata and image properties. For a summary and results of Data+, see this poster. To see the tool in action, see identifier-v3.mp4.
A complete weekly summary of progress on development can be found here.
This effort will augment the LungMAP image annotation tool that allows a user to notate specific anatomical features directly onto an image within the web environment by placing a graphic indicator (arrow, dot or by circling an area of interest within the image) within the image. The tool links the graphic indicator to a selected ontology term to enrich a viewer experience of the image. In later release of the LungMAP website, these annotated images will be highlighted.