Latest question:
How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!




By Lee J. Nelson
Contributing Editor
Drawing on computer graphics and machine vision techniques, the methodology enables detailed three-dimensional CT-based procedure mapping and follow-on image-guided bronchoscopy.
The 3D CT data demarcate the virtual-world representation of the patient's chest while the bronchoscope delivers a continuous two-dimensional video stream of the airway-tree interior. Barrel distortion, inherent in the bronchoscope camera, ensures a wide-angle (fish-eye) vista. Although the deformation yields greater detail near the image center, it nonlinearly stretches the scene. Thus, a corrective transformation must be applied to match the bronchoscope's field-of-view to the virtual-world, as defined by the CT imagery. Such mutual data, which arise in information theory to measure the statistical interrelationship between two random variables, commonly is used to register images from different modalities.
In endoluminar airway images, the significant information resides primarily near airway bifurcations, the corresponding dark "holes" which lead to upcoming airways and near large (darker) airway wall deviations. Specular reflections which appear in small regions of the video images indicate saturated intensity points and are not seen in endoluminal renderings. Drawing upon those observations, the investigators modified their calculations by varying the weights assigned to darker and brighter pixels.
Other weighing schemes are possible. However, by attenuating brighter pixels while emphasizing darker ones, greater emphasis is given to the significant structures.
During bronchoscopy, the system registers and fuses CT data with live video from the bronchoscope. A rendered three-dimensional navigational path to each preplanned biopsy site provides the physician with direct visual guidance on maneuvering the endoscope through major airways and inserting the biopsy needle. In preliminary testing, the system increased physicians' vision of procedural circumstances, simplified decision making and seemed to heighten biopsy success. Notably, operator accuracy improved as differences in skill-levels were mitigated.