Application Of Fractal Texture Analysis To Segmentation Of Dental Radiographs

Fractal Surfaces

“…Bale and Schmidt (1984) have discussed the use ofX-ray scattering to study porous fractals. Other medical and dental radiography applications of fractal geometry have been demonstrated as weH (Lundahl, Ohley et al 1985;Lundahl, Ohley et al 1986;Dellepiane, Serpico et al 1987;Chen, Daponte et al 1989;Kuklinksi, Chandra et al 1989;Caldwell, Stapleton et al 1990;Chuang, Valentino et al 1991;Rigaut 1991;Fortin, Kumaresan et al 1992). Xiao, Chu et al (1992) have used measurement of the fractal dimensions of brightness patterns for recognition in mammograms, and Recht (1993) is doing the same forpap smears.…”

“…Other medical imaging applications (Lundahl, Ohley et al 1985;Lundahl, Ohley et al 1986;Dellepiane, Serpico et al 1987;Chen, Daponte et al 1989;Kuklinksi, Chandra et al 1989;Caldwell, Stapleton et al 1990;Rigaut 1991;Fortin, Kumaresan et al 1992;Xiao, Chu et al 1992) of the fractal measurement ofbrightness patterns from radiography have been used for identification and image segmentation. It appears that fractal geometry appears with some frequency in natural situations, and consequently its use as a characterization tool has diagnostic utility.…”

Fractals in dentistry

“…Ruttimann et al [87] reported a combined in vitro and in vivo study on the possibility of using FD from radiographs to characterize alveolar structure. To compute the FD they used the power spectrum method [88]. They reported a FD increase after induced demineralization for the in vitro study, noting also that FD varied with the location of the ROI, and that his results were the same no matter the projection angles.…”

Fractal analysis of bone images

“…In another technique [14] a two dimensional gray level image is represented as a three-dimensional surface whose height at each point represents the gray level at that point and the surface area is measured at different scales. It has also been shown that [15] an ndimensional fractal object can be characterized by the fractional Brownian motlion of n variables and that the relationship between the power spectral density and r are independent of the projection [23]. This makes the fractal dimension computed from the projections of an n-dimensional fractal object represent the original object.…”