Dictionary learning based superpixels clustering for weakly-supervised semantic segmentation
Peng Ying; Jing Liu; Hanqing Lu
2015
会议名称IEEE International Conference on Image Processing
会议录名称
会议日期September 27-30, 2016
会议地点Quebec City, QC, Canada
摘要
; The task of weakly-supervised semantic segmentation is solved by assigning image-level labels to over-segmented superpixels. Considering that superpixels are geometrically and semantically ambiguous for label assignment, we propose a joint solution of semantic segmentation to enhance the learnability of superpixels. First, our model includes a spectral clustering item and a discriminative clustering item to obtain some clustering subsets of superpixels (ideally semantic regions), which are more separable semantically than independent superpixels. Second, sparse coding based feature for superpixel is adopted to make the representation robust to noise, and the dictionary for the sparse representation is learned together with the above clustering items. Third, a weakly supervised item for superpixels, transferred from image-level labels, is attached. We jointly formulate the above problems as a non-convex objective function, and optimize it by the constraint concave-convex programming (CCCP) algorithm. Extensive experiments on MSRC-21 and LabelMe datasets prove the effectiveness of our approach.
关键词Dictionary Learning Weak Supervision Semantic Segmentation
文献类型会议论文
条目标识符http://ir.ia.ac.cn/handle/173211/11600
专题模式识别国家重点实验室_图像与视频分析
通讯作者Jing Liu
推荐引用方式
GB/T 7714
Peng Ying,Jing Liu,Hanqing Lu. Dictionary learning based superpixels clustering for weakly-supervised semantic segmentation[C],2015.
条目包含的文件 下载所有文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
应鹏_ICIP.pdf(1666KB)会议论文 开放获取CC BY-NC-SA浏览 下载
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
[Peng Ying]的文章
[Jing Liu]的文章
[Hanqing Lu]的文章
百度学术
百度学术中相似的文章
[Peng Ying]的文章
[Jing Liu]的文章
[Hanqing Lu]的文章
必应学术
必应学术中相似的文章
[Peng Ying]的文章
[Jing Liu]的文章
[Hanqing Lu]的文章
相关权益政策
暂无数据
收藏/分享
文件名: 应鹏_ICIP.pdf
格式: Adobe PDF
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。