Ay @ BE ‘Aza mH Ags = Axa @ BS Aga = Ays@ BP “Aggconverted 10 No more changes after this, ™-conneetivity. FIGURE 4-11 (a) Sequence of rotated structuring elements used for thinning. (b) Set A. {¢) Result of thinning with the first element. (¢)-(i) Results of thinning with the next bed OF G seven elements (there was no change between the seventh and eighth elements). fi @) Result of using the first four elements again, (I) Result after convergence. (m) K fm Conversion to m-connectivity 4.8 Image Segmentation Image segmentation divides an image into regions that are connected and have some similarity within the region and some difference between adjacent regions. The goal is usually to find individual objects in an image. For the most part there are fundamentally two kinds of approaches to segmentation: discontinuity and similarity. = Similarity may be due to pixel intensity, color or texture, = Differences are sudden changes (discontinuities) in any of these, but especially sudden changes in intensity along a boundary line, which is called an edge. There are three kinds of discontinuities of intensity: points, lines and edges. The most 9 @ scanned with OKEN Scannera small mask over the image. The mask common way to look for discontinuities is to scan a determines which kind of discontinuity to look for. Only slightly more common than point detection is to find one pixel wide line in an image. For digital images the only three point straight lines are only horizontal, vertical, or diagonal (+ or ~45°). 4.9 Edge Linking & Boundary Detection Two properties of edge points are useful for edge linking: — the strength (or magnitude) of the detected edge points = their directions (determined from gradient directions) + This is usually done in local neighbourhoods. + Adjacent edge points with similar magnitude and direction are linked. + For example, an edge pixel with coordinates (xo,¥0) in a predefined neighbourhood of (x,y) is similarto the pixel at (x) if [XC 9) - VC. %9)| SE, Esa nonnegative threshold nonegative angle threshold lax. y)—a(x, ¥)|< A, A? Hough transform: a way of finding edge points in an image that lie alonga straight line. Example: xy-plane v.s. ab-plane (parameter space) viz axj +b + The Hough transform consists of finding ll pairs of values of @ and p which satisfy the equations that pass through (x,y). + These are accumulated in what is basically a 2-dimensional histogram, + When plotted these pairs of @ and p will look like a sine wave. The process is repeated forall appropriate (x,y) locations. 4.10 Thresholding Global ~ T depends only on gray level values Local—T depends on both gray level values and local property Dynamic or Adaptive ~ T depends on spatial coordinates Different approaches possible in Graylevel threshold + Interactive threshold 10 @ scanned with OKEN Scanner+ Adaptive threshold + Minimisation method en Fig. 4.11 Gray level thresholding 4.11 Region based Segmentation + Edges and thresholds sometimes do not give good results for segmentation. + Region-based segmentation is based on the connectivity of similar pixels in a region. — Each region must be uniform. = Connectivity of the pixels within the region is very important. + There are two main approaches to region-based segmentation: region growingand region splitting. Basic Formulation + Let R represent the entire image region. For example: P(Ri)=TRUE if all pixels in Ry have the same gray level. Region splitting is the opposite of region growing. — First there is a large region (possible the entire image). — Then a predicate (measurement) is used to determine if the region is uniform. — Ifnot, then the method requires that the region be split into two regions. — Then each of these two regions is independently tested by the predicate (measurement). — This procedure continues until all resulting regions are uniform. The main problem with region splitting is determining where to split a region. One method to divide a region is to use a quad tree structure. Quadtree: a tree in which nodes have exactly four descendants, The split and merge procedure: — Split into four disjoint quadrants any region R, for which P(Ri) =FALSE. — Merge any adjacent regions R,and Rj for which P(RUR) = TRUE. (the ul @ scanned with OKEN Scanner