SR-clustering: Semantic regularized clustering for egocentric photo streams segmentation

详细信息    查看全文

• 作者：Mariella Dimiccoli ; ¹ ; ^a ; ^c ; ^{mariella.dimiccoli@cvc.uab.es}Author Vitae ; Marc Bolañ ; os¹ ; ^a ; ^{marc.bolanos@ub.edu}Author Vitae ; Estefania Talavera^a ; ^b ; ^{etalavera@ub.edu}Author Vitae ; Maedeh Aghaei^a ; ^{maghaeigavari@ub.edu}Author Vitae ; Stavri G. Nikolov^d ; ^{stavri.nikolov@imagga.com}Author Vitae ; Petia Radeva ; ^a ; ^c ; ^{petia.ivanova@ub.edu}Author Vitae
• 关键词：Temporal segmentation ; Egocentric vision ; Photo streams clustering
• 刊名：Computer Vision and Image Understanding
• 出版年：2017
• 出版时间：February 2017
• 年：2017
• 卷：155
• 期：Complete
• 页码：55-69
• 全文大小：5344 K
• 卷排序：155

文摘

A framework for semantic egocentric photo streams segmentation is presented. Contextual and semantic information is extracted by employing a CNN approach. A vocabulary of concepts is defined by relying on linguistic information. Images sharing contextual and semantic attributes are grouped together. Our framework is suited for event recognition, semantic indexing and summarization.