Abstract:After introducing several results relating to the modification of the homotopy of gradient functions based on extrema in the base image and building on earlier results in morphological scale-space, we introduce a scale-space monotonicity theorem for regions of an image defined by watersheds of a gradient function modified to retain only the local minima or maxima of its smoothed parent image. We then illustrate the theorem with an example of the scale-space extraction of texture features from the nuclei of cer… Show more
“…Thus, after cleanup, edge (3, 4, 5) has been removed, being an internal edge in the tree rooted at 3. Furthermore, edge (2,4,8) and edge (2, 3, 7) connect the same distinct trees rooted at 2 and 3. Consequently, the lighter edge, (2,3,7), is retained in the graph, while edge (2,4,8) is elided.…”
Section: Description Of the Parallel Algorithmmentioning
confidence: 99%
“…For the image example in Fig. 1a [5,12], [6,13]], bcg 2 =[ [9,0], [7,0], [10,13], [8,1]], bcg 3 =[ [11,0], [12,5], [13,6], [13,10]]. The local WNG, wng, is locally computed in each processor, as described in Section 2.…”
Section: Description Of the Parallel Algorithmmentioning
confidence: 99%
“…Suppose that in Fig. 1b pixels (3, 3), (4,6), and (8,6) ((line, col ) coordinates are relative to the (0, 0) top left corner of the image) are marked; then regions 1 and 4 have to be appended to a marked area, labeled 2 and 3. Let us first remark that merging is performed regionwise, and second that a catchment basin which contains more than one marked pixel is not further split (see region labeled 3 in Fig.…”
“…Thus, after cleanup, edge (3, 4, 5) has been removed, being an internal edge in the tree rooted at 3. Furthermore, edge (2,4,8) and edge (2, 3, 7) connect the same distinct trees rooted at 2 and 3. Consequently, the lighter edge, (2,3,7), is retained in the graph, while edge (2,4,8) is elided.…”
Section: Description Of the Parallel Algorithmmentioning
confidence: 99%
“…For the image example in Fig. 1a [5,12], [6,13]], bcg 2 =[ [9,0], [7,0], [10,13], [8,1]], bcg 3 =[ [11,0], [12,5], [13,6], [13,10]]. The local WNG, wng, is locally computed in each processor, as described in Section 2.…”
Section: Description Of the Parallel Algorithmmentioning
confidence: 99%
“…Suppose that in Fig. 1b pixels (3, 3), (4,6), and (8,6) ((line, col ) coordinates are relative to the (0, 0) top left corner of the image) are marked; then regions 1 and 4 have to be appended to a marked area, labeled 2 and 3. Let us first remark that merging is performed regionwise, and second that a catchment basin which contains more than one marked pixel is not further split (see region labeled 3 in Fig.…”
“…In non-homogeneous or noise embedded images there is not a one to one relation between regional minima and objects of interest. This results in an over segmentation in the majority of images, in other words, after WT each of the objects is represented by more than one region [17][18][19] [26] [30]. To avoid this over segmentation we resort to the selection of a single marker for each object of interest.…”
“…The watershed transform based segmentation approach works on morphological principles [22][23][24][25]. If we regard a grayscale image as a topographic relief, the gray value at a given location represents the elevation at that point.…”
The wavelet transform as an important multi resolution analysis tool has already been commonly applied to texture analysis and classification. Mathematical morphology is very attractive for automatic image segmentation because it efficiently deals with geometrical descriptions such as size, area, shape, or connectivity that can be considered as segmentation-oriented features. This paper presents an image-segmentation system based on some well-known strategies implemented in a different methodology. The segmentation process is divided into three basic steps, namely: texture gradient extraction, marker extraction, and boundary decision. Texture information and its gradient are extracted using the decimated form of a complex wavelet packet transform. A novel marker location algorithm is subsequently used to locate significant homogeneous textured or non textured regions. The goal of boundary decision is to precisely locate the boundary of regions detected by the marker extraction. This decision is based on a region-growing algorithm which is a modified flooding based watershed algorithm.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.