Segmentation of road scenes using color and fractal-based texture classification

Presents a knowledge-based method for segmenting color images of road scenes. The knowledge base contains color and texture representations of objects typically found in such scenes. The segmentation algorithm uses the knowledge base to classify each image region according to its chromaticity and fractal dimension. A key feature of the texture classifier is its robustness with respect to image scale and contrast. We also discuss extensions of our method to other applications