Efficacy of Image Processing Filters in the Detection of Proximal Caries in Digital Bitewing Radiograph

Background and objectives: recently, most digital imaging systems provide the option of image post-processing with different techniques. the aim of present study was to compare the performance of digital radiography with and without image processing filters in the detection of proximal recurrent caries.subjects and methods: in this in vitro study,amalgam class 2 cavities were prepared on both proximal surfaces of 52 healthy premolar teeth. artificial caries lesions were made by a round 0.5 mm bur in mesial or distal of the boxes randomly. standardized digital radiographs were acquired using the digora® optime system. unfiltered and filtered (low sharpen, intermediate sharpen, high sharpen and inversion) images were assessed by 2 independent observers.results: sensitivity, specificity and overall accuracy evaluated with sub curve analysis of roc (az), anova/tukey and spearman rank correlation coefficient tests. the sensitivity, specificity and overall accuracy of the low sharpen, unfiltered images and inversion was significantly higher than intermediate sharpen and high sharpen filter images (p?0.05). the low sharpen filter images had the highest sensitivity, specificity and overall accuracy. the high sharpen filter images had the lowest sensitivity, specificity and overall accuracy.conclusion: the low sharpen filter may be an acceptable processing filter for the detection of proximal recurrent caries. the intermediate sharpen and high sharpen filters had the worst performance than the other modalities and can not be recommended for the detection of proximal recurrent caries.