MAXILLARY ALVEOLAR BONE IMAGING - WIENER FILTER DESIGN

Optimal digital filter design is essential for noise suppression in the detection of subtle radiographic bony changes. The purpose of this study was to determine the signal (Ps) and noise (Pn) power spectra of sampled maxillary alveolar bone, to derive the optimal Wiener filter transfer function, H, from sets of Ps and Pn, and to quantify noise suppression through application of this filter. Sixteen standardized radiographs were made of five interproximal bony areas, each area from a separate dry human maxilla. The radiographs were digitized (0.02 mm/pixel resolution) and identical profiles (scan lines) generated between lead markers for each set of films. Ps was calculated from the averaged scan line of each set, Pn was calculated from the difference between the noisy images and the averaged scan line for each set, and an average maxillary H was calculated from the sets of Ps and Pn. Filtering of the 80 noisy sample radiographic profiles with H resulted in a 39% reduction in noise. Application of this digital filter should significantly improve detection of radiographic bony changes in the maxilla.