COMPENSATION FOR THE SIGNAL-PROCESSING CHARACTERISTICS OF ULTRASOUND B-MODE SCANNERS IN ADAPTIVE SPECKLE REDUCTION

A systematic method to compensate for nonlinear amplification of indiv idual ultrasound B-scanners has been investigated in order to optimise performance of an adaptive speckle reduction (ASR) filter for a wide range of clinical ultrasonic imaging equipment. Three potential method s have been investigated: (1) a method involving an appropriate select ion of the speckle recognition feature was successful when the scanner signal processing executes simple logarithmic compression; (2) an inv erse transform (decompression) of the B-mode image was effective in co rrecting for the measured characteristics of image data compression wh en the algorithm was implemented in full floating point arithmetic; (3 ) characterising the behaviour of the statistical speckle recognition feature under conditions of speckle noise was found to be the method o f choice for implementation of the adaptive speckle reduction algorith m in limited precision integer arithmetic. In this example, the statis tical features of variance and mean were investigated. The third metho d may be implemented on commercially available fast image processing h ardware and is also better suited for transfer into dedicated hardware to facilitate real-time adaptive speckle reduction. A systematic meth od is described for obtaining ASR calibration data from B-mode images of a speckle producing phantom.