Providing clinicians and basic scientists with knowledge representation tools built on the need to visually identify and label features on an image and add expert knowledge for collaboration and sharing will improve the process of clinical and scientific discoveries.
In the biomedical-imaging environment, images are generated from multiple imaging modalities that include clinical photos, radiographs, histology, and microscopy. At each stage of the clinical management of a clinical or research study, multiple images are acquired from multiple imaging modalities. Furthermore, each image is interpreted by one or more experts and generally, must be shared by one or more experts.
One challenge among many, is the need to incrementally link related information to images and use that same information to link images with other images from the same imaging study that in turn supports and facilitates retrieval and re-use of the imaging knowledge base. The typical process is for each expert to interpret images independently from other experts without the ability to link their results.
The solution is one that facilitates the incremental collection of expert knowledge in the form of visual annotations that can be applied to images in a non-destructive manner thereby making the collected information available to other experts. The technology that supports developing such a solution could potentially reduce repetition of work and image re-interpretation by allowing incremental addition of expert knowledge in the form of non-destructive visual annotations with related textual descriptions, all without destroying the image. Such a solution could support multi-specialty authoring of images resulting from research studies and reduce the replication of images by storing the annotated information linked to the image(s) according to the domain expert that created them.

The figures demonstrate how visual annotations can be applied to images from different imaging modalities without destroying the underlying image.
Any given solution must provide the ability to add or incorporate a standard lexicon or vocabulary to a single annotation, a set of annotated information, the image, or the entire session and export or save the annotation session in a structured output format. Adopting a structured output format of the image, visual annotations, and textual information in a format such as eXtensible Markup Language (XML), simplifies storage and retrieval by facilitating indexing and cataloging of the knowledge base. A solution that is designed to maintain the knowledge base of images and visual annotations has broad application in both the clinical and scientific community.
Comments (1)
Linking information in an iterative manner to images will allow for a collection of information for reference. Great article!
Posted by Sandra Armstrong | October 2, 2007 8:06 AM
Posted on October 2, 2007 08:06