Target segmentation and detection based on fractal dimension feature

A modified fractal dimension algorithm is presented. By dividing each pixel into a grid of H × H, the image can be viewed as a 2D dot set. The value of the pixel equals to the number of the dots within the grid. The definition of the box dimension can be followed in 2D graph. The FD of each pixel for the high-pass image is computed as its fractal feature. A number of different type textures in the background are similar in feature value. The feature values of target areas are higher than that of background. The targets are detected by segmenting and clustering recursively. The thresholds decrease from the biggest feature value. When the number of clusters does not change after several iterations, the clusters are taken as the detection result.