Comparison of the diagnostic accuracy of different image processing algorithms in the diagnosis of jawbone defects

Abstract

Background and Aim: Intra-bone lesions are common in lesions, among them periapical lesions have a higher incidence. Todays, all direct digital imaging systems provide various types of image processing techniques. The purpose of this study is to investigate the effects of different digital image processing techniques on the diagnosis of bony defects. Materials and Methods: In this experimental study, 50 sheep inferior half-jawbones were used. Lesions were created by round bars with numbers 1/4, 1/2, 1, 2 and 4. Digital images were obtained by edge enhance, sharpening, invert gray scale processing method and unprocessed investigated by two radiologists and in case of disagreement them observers by third observer. The method of recording the results was definitive diagnosis of lesion number 1, misdiagnosis number 2 and possible diagnosis number 3 .The results were analyzed by chi-square and Fisher's exact tests. Results: Percent of diagnostic sensitivity in unprocessed Images (71.8%) and images processed with Sharpen filter (71.4%) were significantly higher than images processed with invert grayscale (68.5%) and edge Enhancer (53%).(p<0.001) Conclusion: To detect bone lesions, unprocessed and processed digital images with Sharpen filters have higher diagnostic accuracy.

Keywords

• digital radiography
• image processing
• bony defect