This paper is concerned with the development of a computational methodology based on fractal geometry for determining 3D structure of protein with imagery projection operations. In this investigation, the density-map image is a 2D projection of the 3D electron density map according to its depth of the density distribution along the projection direction. We extract fractal features of the density-map image in a region and use these features to look for candidate regions with similar patterns of density-map. We analyze its fractal signatures for determining 3D pattern of regions of density distribution. This contribution presents preliminary results of such a study, wherein the protein surface was assumed to be a fractal, and computed fractal feature (fractal dimension and fractal signature) were analyzed and found to possess fairly reasonable pattern for improving the discrimination abilities of the protein structure. © 2002 IEEE.