Watershed color image segmentation based on prior knowledge
Liu, Jialu; Yang, Minghui; Peng, Silong,; Liu JL(刘佳璐)
2006
发表期刊Computer Engineering
卷号32(21)期号:2006年11期页码:pp 28-30 (EI)
摘要This paper proposes a new image segmentation algorithm based on watershed transformation combined color information and uses Bayesian inference on this watershed image.It transforms RGB color image to LUV space,uses watershed transformation based this color image.This paper calculates the energy of the label image result from the color image watershed transformation by designing a prior density that penalizes the area of homogeneous parts in images.The segmentation problem is the maximizing a posteriori estimation of the set of object area result from the watershed labeled.Then could find the optimal area of object,and the other area of the image looked upon background area.This algorithm not only solves the over-segmentation problems of watershed transformation,but also uses color information and prior knowledge.The experiments indicate the algorithm is effective for image segmentation.
关键词Bayesian Framework / luv Color Space / watershed Transformation / map Estimation
文献类型期刊论文
条目标识符http://ir.ia.ac.cn/handle/173211/12978
专题智能制造技术与系统研究中心_多维数据分析
通讯作者Liu JL(刘佳璐)
推荐引用方式
GB/T 7714
Liu, Jialu,Yang, Minghui,Peng, Silong,,et al. Watershed color image segmentation based on prior knowledge[J]. Computer Engineering,2006,32(21)(2006年11期):pp 28-30 (EI).
APA Liu, Jialu,Yang, Minghui,Peng, Silong,,&刘佳璐.(2006).Watershed color image segmentation based on prior knowledge.Computer Engineering,32(21)(2006年11期),pp 28-30 (EI).
MLA Liu, Jialu,et al."Watershed color image segmentation based on prior knowledge".Computer Engineering 32(21).2006年11期(2006):pp 28-30 (EI).
条目包含的文件
条目无相关文件。
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[Liu, Jialu]的文章
[Yang, Minghui]的文章
[Peng, Silong,]的文章
百度学术
百度学术中相似的文章
[Liu, Jialu]的文章
[Yang, Minghui]的文章
[Peng, Silong,]的文章
必应学术
必应学术中相似的文章
[Liu, Jialu]的文章
[Yang, Minghui]的文章
[Peng, Silong,]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。