EDGE-DETECTION IN UNTEXTURED AND TEXTURED IMAGES - A COMMON COMPUTATIONAL FRAMEWORK

In this paper, a unified computational framework is presented for faci litating edge detection both in untextured as well as textured two-dim ensional (2-D) images. The framework is based on a complete set of dif ference operators which are easily configurable from a set of orthogon al polynomials, The widely known Roberts, Sobel, Prewitt, and Marr's L OG edge operators can easily be represented in terms of these operator s, For detection of untextured or textured edges, the proposed operato rs are employed to separate out the responses toward edge or texture a nd noise, Untextured edges are detected by maximizing signal-to-noise ratio (SNR) or identifying the zero crossings in the second directiona l derivatives. Textured edges are detected in two stages, First, the s ignificance of responses toward texture is computed statistically in o rder to test the presence of microtexture and compute a local descript or called ''pronum'' for its representation, Finally, a global descrip tor for texture called ''prospectrum'' is obtained by observing the fr equency of occurrence of pronums. The textured edges are detected at t he second stage by applying the methods of detection of untextured edg es on these prospectrums. The results are encouraging.