Our Algorithmics division uses computer vision technology to develop algorithmic engines that automate many of the tasks routinely performed by clinicians interacting with volumetric imaging data. These engines are integrated into the product lines of medical device companies developing end user clinical solutions. We have an outstanding track record of timely solving the most difficult algorithmic challenges, and delivering robust, clinically-ready algorithms to our partners.
Our team is composed of highly experienced algorithmic engineers, and at last count we had a total of over 75 years of combined experience developing volumetric medical imaging algorithms. Our team members' knowledge covers many key technologies including segmentation, registration, rendering, automated detection (CAD), vessel analysis and even volumetric CT reconstruction. Our engineers led many industry innovations and hold more than 25 patents issued or pending. We like to think of them as the industry "A-team" (where the A stands for either automation or algorithmic excellence).
Our team typically uses our WithinsightTM Framework (WIF) to develop the algorithmic engines, enabling us to create more sophisticated, versatile and robust algorithms will less effort. The types of tasks performed by Withinsight “engines” include:
Fully automated segmentation: Tagging of regions in volumetric (3D and 4D) data with anatomical labels (tissues and organs). Adding geometrical descriptors, such as centerlines, branching trees, lumen boundaries, anatomical landmarks, analytical surfaces, etc.
Elastic registration: Automatically matching homologous locations in multiple scans of the same patient, or even different patients, taken at different times by the same, or even different, modality. Quantifying and predicting change in anatomic regions over time.
Abnormality detection (CAD): Detection and marking of suspected anatomical abnormalities (eg, lesions), as well as providing quantitative statistical measurements, such as diameters, size and density, of each abnormality.
Volume visualization: Generation of informative views from volumetric data. Exposing otherwise hidden structures by using transparency effects and “flattening” of complex structures.
We support the integration of our algorithmic engines with our partners’ applications, ensuring the most effective and efficient interaction between application workflow and the algorithmic engine. We support our partners’ product deployment through Beta phases and initial clinical use, and provide a simple and effective approach to project delivery by milestones.
