Real-time equalization of region-of-interest fluoroscopic images using binary masks

In region-of-interest (ROI) radiologic imaging, the x-ray beam is attenuated peripherally to the region of interest to reduce patient exposure. This attenuation reduces the peripheral image brightness which may cause contrast in the periphery to also be reduced due to video chain nonlinearity. For optimal viewing, it is necessary that the image brightness and contrast in the periphery be brought back to the levels in the ROI. Previously, digital subtraction angiography roadmapping equipment has been used for this equalization; however, the procedure is not independent of patient and gantry motion. A new motion independent method to achieve this equalization involves dividing the real-time video signal into two digital streams one of which is brightness and contrast enhanced. A pre-acquired binary mask image is created by thresholding the image of a uniform object obtained with the ROI filter in place. This binary mask is used to control the recombination of the two image streams in a digital pipeline processor in order to select the ROI from the unprocessed stream and the periphery from the enhanced stream. This system provides image equalization at 30 frame/s for real-time ROI imaging display. Images from this method demonstrate excellent image quality even for peripheral exposure reduction factors exceeding 10.