The perceptibility curve test applied to CCD and two methods of digitization of dental film-based radiographs

Objectives: Several methods of image acquisition are accessible in dentistry. There is no overall acceptable method for image digitization so all different types of images can be comparable. The objective of this study was to compare the diagnostic accuracy of different methods of image digitization. Methods: This accuracy diagnostic test study used perceptibility curve test which first introduced by de Balder was applied. In this test a test object is used which is usually made by aluminum. Different levels of thickness and number of holes in the test object were necessary to have different levels of contrast. Images from film and CCD and digitized images by means of CCD scanners and digital camera were prepared. Nine observers assessed the images. Data collected was delivered to SPSS 13 software and for each image acquisition method, interclass correlation coefficience was computed and compared to the gold standard. Results: Mean sensitivity, specifity, positive and negative like hood ratios in dependence on material thickness and the background gray value were calculated. In regions of high optical density the sensitivity for the film images was highest (0.994) following by CCD (0.905), scanner (0.889) and camera (0.821). Difference between CCD images and scanner images was not significant. In dark regions of no dark holes the sensitivity was highest for film images (0.832) following by CCD (0.798), camera (0.714) and scanner (0.615) Difference between film and CCD images was not significant. Conclusion: The diagnostic quality of radiographic films was better than digital CCD sensors. For digitizing analog images scanners were better than digital cameras.