Abstract | ||
---|---|---|
This paper presents a scheme for color image segmentationbased on watershed algorithm that solves over-segmentationproblem. A lexicographical order is adopted on HSI(Hue-Saturation-Intensity) color model so that morphologicaloperations can be applied on color images directly. A markerimage is then defined using both chromaticity and intensityinformation of the image. Together with multi-scale gradientimage and marker image, watershed segmentation can makesure to partition image into meaningful objects and avoidfurther segmentation of homogeneous regions.The results of the proposed algorithm are compared withthose of other standard methods. Preliminary experimentshave shown a better result and indeed solvedover-segmentation problem. |
Year | DOI | Venue |
---|---|---|
2004 | 10.1109/MULMM.2004.1264990 | MMM |
Keywords | Field | DocType |
marker image,avoidfurther segmentation,color model,lexicographical order,color image,better result,watershed algorithm,watershed segmentation,proposed algorithm,homogeneous region,lexicographic order,image segmentation,mathematical morphology | Computer vision,Image gradient,Scale-space segmentation,Color histogram,Pattern recognition,Image texture,Computer science,Binary image,Segmentation-based object categorization,Image segmentation,Artificial intelligence,Color image | Conference |
ISBN | Citations | PageRank |
0-7695-2084-7 | 0 | 0.34 |
References | Authors | |
7 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Shwu-huey Yen | 1 | 42 | 9.07 |
An-Chi Tai | 2 | 0 | 0.34 |
Chia-Jen Wang | 3 | 17 | 3.25 |