Abstract
The subtracted images obtained from double arterial hepatic dynamic CT using multi-detector row CT(MDCT) are useful in detecting hepatic tumors which show subtle enhancement in the second phase. However, manual subtraction of MDCT images is time-consuming for radiologists. So, we have developed a fully automated dynamic subtraction algorithm of hepatic CT by use of 3D phase only correlation method for improvement of detection of hepatic tumors. Our algorithm was as follows. In the first step, global-matching processing was performed based on extracted bone regions. In the next step, local-matching processing was performed for the liver area. In the final step, subtraction processing was performed. We evaluated our algorithm using 12 clinical cases, and obtained promising results.
