Category-specific object segmentation via unsupervised discriminant shape

详细信息    查看全文

• 作者：Lingzheng Dai ^{lingzhengdai@163.com}Author Vitae ; Jian Yang ; ^{csjyang@njust.edu.cn}Author Vitae ; Liang Chen ^{chinandy@foxmail.com}Author Vitae ; Junxia Li ^{junxiali99@163.com}Author Vitae
• 关键词：Object segmentation ; Unsupervised discriminant clustering ; Graph cut
• 刊名：Pattern Recognition
• 出版年：2017
• 出版时间：April 2017
• 年：2017
• 卷：64
• 期：Complete
• 页码：202-214
• 全文大小：1559 K
• 卷排序：64

文摘

A novel unsupervised discriminant shape is proposed for segmenting images. An unsupervised discriminant clustering method is built for precise shape prior. Our model can produce reliable segmentation for category-specific objects. Evaluations on 4 datasets demonstrate the effectiveness of our method.