Digital gray-level transformation for the reduction of redundant shadows in rotational panoramic radiography

In rotational panoramic radiography, the tomographic motion in combination with a slit scanning method make an image layer wide. But some objects outside the image layer are hard to blur and they result to redundant shadows. The gray-level transformation, as one form of digital image processing for the reduction of these redundant shadows, was evaluated. Two methods, the gamma transformation and the histogram flattening method, were examined. A drum scanner was used as the image scanner. The panoramic image on the x-ray film was turned into an 8 bit digital image on the image memory, which had the size of a 512 x 480 matrix. In rotational panoramic radiography, there are low density and contrast regions where redundant shadows of the cervical vertebrae and the mandibular ramus are superimposed on the tomographic image. The histogram of the gray-level was suppressed for the lower gray-levels. The stretching of this gray-level distribution was effective in the reducing redundant shadows. When processed by gamma transformation, the smaller gamma coefficient below 1.0 clarified the tomographic image, and when processed by the histogram flattening method, the setting of the level (L) value around 64-128 effectively reduced the redundant shadows. However, the effectiveness of both gray-level transformations was greatest in restricted cases in which the area where redundant shadows were superimposed on the tomographic image was comparatively large.