Significant Region Detector

Technology #2008-027

An algorithm has been developed to computationally define significance (in the sense of objects in noisy images which differ significantly from the background noise) in two and three dimensional image volumes. The algorithm uses this definition of significance in the course of reviewing image volumes containing zero mean noise and optional object regions, to automatically and reproducibly identify the object regions, and to do so with very high sensitivity and specificity. The algorithm is broadly applicable to image processing both within various sub-fields of medical imaging, and within other imaging disciplines (e.g. remote sensing, surveillance, etc.)