Morphological based technique for image segmentation
M.S.Z. Sarker, Wooi-Haw Tan, Rajasvaran Logeswaran · 2008
In this paper, a morphological image segmentation technique based on the watershed algorithm is proposed. New morphological based pre-and post-processing techniques are proposed to reduce over-segmentation, by means of merging and removing spurious segments. The preprocessing aims at removing trivial regions as well as background noise. The post-processing produces a more concise region representation of the watershed-segmented image, where a region adjacency list (RAL) is built for the region merging process. To control the merging, a similarity function is defined, whence the most similar neighbouring regions are merged. The proposed technique produces effective and significant results in successfully segmenting various objects when tested on a series of well known test images, as shown in this paper