Sharpening based image enhancement algorithms in reducing the disagreement of medical images subjective evaluation

Subjective evaluation of the abnormalities in dental images faces a few challenges such as low contrast. The problem arises due to the fix regulation of small X-ray dosage. Thus there are applications of image enhancement on the dental images and it is an acceptable technique to improve image quality and better diagnosis. Low contrast images could hinder the subjective evaluation and contribute to disagreement between the evaluators. This work investigates the performance of original and enhanced dental images towards the disagreement issues of evaluating image quality and detecting dental abnormalities. The abnormalities of interest are periapical radiolucency (PA), widen periodontal ligament space (widen PDLs) and loss of lamina dura (Loss of LD). The work begins with collecting the raw intra-oral dental images. Then sharpening based contrast image enhancement algorithms were applied to the images. After that, the images were evaluated by dentists towards the image quality and the abnormalities mentioned. The disagreements among the evaluators were determined using standard deviation formula. Results show that disagreement issue among the evaluators do exists to the extent of above 90%. Comparing to the original images show that enhanced images are able to slightly reduced the subjective evaluation disagreement in image quality and abnormalities.