Comments on noise and resolution of the DenOptix radiography system

Relatively low signal/noise (S/N) ratios, which are not substantiated by the results reported here, have been reported for the DenOptix digital intraoral radiography system. Unexplained poor performance has also been reported for imaging of the gingival tissue. I sought to discover the probable sources of the discrepancies and to present data on the effects of pixel size and screen type on noise and resolution. Methods. S/N ratios were measured with an aluminum phantom, and noise was measured with a step-wedge. The radiographs were scanned with both default and special, manufacturer-supplied software to obtain 8-bit rescaled data and 16-bit raw data. The data were compared with previously published data for the DenOptix and Digora systems.

The S/N ratios of up to 17 obtained here were much higher than previously reported. Previously reported low ratios and poor performance for the imaging of gingival tissue were likely caused by digital clipping of 8-bit images, which should not affect bony tissue images in normal clinical radiographs. With nonstandard white phosphor screens, S/N ratios of up to 18 were observed, with only a minor loss of resolution. For the smallest pixel size (42 micro m), resolutions of > or =10 line-pairs/mm and > or =8 line-pairs/mm were observed for standard blue and white phosphor screen types. If the smallest pixel size is selected, the noise increases by only 25%, which is much less than the theoretical increase.

In terms of both noise and resolution, the DenOptix system compares favorably with the Digora system. S/N ratios for the DenOptix system are comparable with the best of 6 previously tested systems. With the DenOptix system, the smallest pixel size produces both the best resolution and the best noise characteristics. The selection of 16-bit images prevents digital clipping of 8-bit images.