Multiscale denoising and enhancement of 3D rotational X-ray imaging for percutaneous vertebroplasty

3D rotational X-ray (3DRX) imaging has been suggested as an alternative to CT scans for evaluating percutaneous vertebroplasty procedures. We present in this study preliminary work on denoising and enhancement of 3DRX data sets with dyadic wavelet thresholding. Thresholding operators were tuned to accommodate for spatial variability of the anatomical features and the noise energy at each level of decomposition. Results are presented for three clinical data sets with quantitative measurements of contrast to noise ratio (CNR) and qualitative evaluation with three- dimensional rendering of vertebral bodies having cement fillings. Wavelet denoising demonstrated the efficient on enhancement of subtle vertebral edges and clearer delineation of the cement filling contours which improves the accuracy in clinical practice and helps to make 3DRX more competitive when compared to CT.