Abstract
1746
Objectives Frequently in dynamic nuclear medicine imaging studies patient movement can have a serious impact on derived numerical indices e.g. renal transit times. The objective of this study was to investigate how the image projection method for dynamic image registration may be made more robust with respect to both global changes in image mean such as occur in dynamic parathyroid imaging and focal image changes e.g. the gradual appearance of urinary activity in bladder when registering renal studies.
Methods A computationally efficient method for dynamic image registration in nuclear medicine can be achieved by transforming each image in the sequence of frames into two vector projections formed by accumulating pixel values along the rows and columns of the image. The vector projections corresponding to successive frames are in turn used to estimate the individual horizontal and vertical components of any interframe shift by means of a one-dimensional (1-D) PCC-based estimator. The phase correlation method provides a distinct sharp cross correlation peak and furthermore, it is robust to those types of noise that are correlated to the image function, e.g., uniform global variations and gradual local offsets in average intensity. We implement PCC via the Hartley transform which resembles a Fourier transform but is free from the need to process complex numbers leading a faster implementation
Results Hartley transform based PCC was applied to clinical nuclear medicine data including gall bladder imaging , adult and paediatric renal studies all manifesting significant patient movement artifacts. The PCC approach demonstrated robust image registration in all cases including those cases where significant changes in image distribution were present.
Conclusions Hartley transform based PCC is a robust and efficient approach to dynamic image registration in the clinical nuclear medicine environment.