2024-03-28T12:14:12Z
https://meral.edu.mm/oai
oai:meral.edu.mm:recid/4411
2021-12-13T04:04:39Z
1582963302567:1597824273898
user-ucsy
Automatic Natural Image Segmentation by Using MarkerControlled Watershed Method and Region Merging Method
Mon, Khin Lay
Thaw, Mie Mie
Fully automatic image segmentation is adifficult task for natural images because of manyvariations ascontrast and complex background.Conventional segmentation methods require aconsiderable amount of interactive guidance by theuser to attain satisfactory results. Moreover, the mostsubsequent tasks as object detection and imageanalyzing application highlydepend on the accurateand useful segmented result. Therefore, in this paper,an automatic image segmentation method for naturalimages is proposed. The proposed system includesthree approaches: gradient computation with themodified LoG edge filter, marker-controlled watershedsegmentation(MCWS) with automatically markerselection and region mergingapproach that is based onedge strength and homogeneous intensity. The systemcan not only efficiently reduce the significant oversegmentation problem of watershed algorithm and butalso produce the correct and meaningful segmentedimagesIt purposes better performance of segmentedimages for image annotation, objects detection, imageanalyzing task and computer vision.
2015-02-05
http://hdl.handle.net/20.500.12678/0000004411
https://meral.edu.mm/records/4411