Image Processing Techniques

Once a radiograph has been processed, the image is permanent and further adjustments cannot be made. If the image is too dark or too light, the image has to be repeated. However, this is not the case with digital images. All digital systems employ a stable electronic circuit called a bit, or binary digit. A circuit containing a bit can electronically be switched into two states, off or on. Off is represented by a zero and on is represented by a one. If a shade or color is assigned to the zero and the one then only two colors can be used, black or white. Digital devices used in radiographic imaging must be able to represent more than two colors. To image several shades of gray there must be more than one bit, or multibits.

The number of bits corresponds to the number of gray levels dis-played by a particular system and is calculated as follows: L = 2n where L is equal to the number of gray levels and n the number of bits. For example, an 8-bit unit can display 28 or 256 shades of gray in an image. Since a digital image is made up of pixels, each pixel is assigned a numerical value corresponding to a shade of gray, thus the density and contrast of the image is adjusted by varying the numerical values of each pixel. Human vision can differentiate approximately 32 gray levels, which means that the dynamic range of the X-ray detection system and the human eye do not match. As a result, the computer must be manipulated to show the proper density and contrast of the final image. Most manufacturers treat the raw data with a firmware before the image is displayed. This simply means that the software in the system uses certain algorithms or mathematical computations set by the manufacturer to optimize the image. However, once the image is dis-played, it can be further processed by the operator to change parameters as desired.

The factors controlling the dynamic range are the window level and the window width. The window level is the level within the possible shades that is used to create the middle density in the image. This means the window level controls the image density. The window width is the range of gray shades that will be used in creating the image and therefore controls the contrast of the digital image. A computer usually uses 256 shades of gray to display the image. An increase in window width means that more shades of gray are displayed in the image, resulting in a decrease in image contrast. A too narrow or too wide window width can cause information to be missed in the resulting image. When the entire range of densities is displayed, the image will have lower contrast, or more shades of gray. When a smaller range of densities is displayed, the image will have higher contrast, or fewer shades of grays.

Most software programs are equipped with multiple processing tools and filters, but the most widely used are brightness and contrast adjustments. Many of the image processing tools, such as color conversion and three-dimensional filtering, have no known diagnostic value. Since a digital image can be processed, more sophisticated processing methods can be used. Among these methods are digital subtraction, image synthesis, image restoration, and image analysis. The potential of digital imaging in dentistry lies in the development of practical techniques to perform digital subtraction, Tuned Aperture Computed Tomography (TACT), fractal analysis, and the creation of decision support systems.