Абстрактный

Review on Image Segmentation, Clustering and Boundary Encoding

Shah Nilima , PatelDhanesh , JivaniAnjali

Identification of an object in an image is the main objective of image segmentation which can be further applied to object recognition. Image segmentation can be described as a process of segregating an image into different parts which give a meaningful representation to the image. Clustering algorithms are very popular in image segmentation,but the number of regions of the image and initial cluster centroids has to be known a priori. To represent objects, boundary encoding of the objects or segments is also one of the important tasks. Presently,to find the best polygon that achieves a small distortion for a given number of vertices for boundary encoding of segments or objects is a challenging task. Several methods for encoding of object boundaries are known that use the concept of chain codes, B-spline, convex hull or Delaunay triangulation approach etc. A review of various segmentation approaches, clustering algorithms applied to image segmentation as well as boundary encoding algorithms is done in this paper