Show simple item record

dc.contributor.authorKampffmeyer, Michael C.
dc.contributor.authorDong, Nanqing
dc.contributor.authorLiang, Xiaodan
dc.contributor.authorZhang, Yujia
dc.contributor.authorXing, Eric P.
dc.date.accessioned2019-04-16T07:33:34Z
dc.date.available2019-04-16T07:33:34Z
dc.date.issued2018-12-14
dc.description.abstractSalient segmentation aims to segment out attention-grabbing regions, a critical yet challenging task and the foundation of many high-level computer vision applications. It requires semantic-aware grouping of pixels into salient regions and benefits from the utilization of global multi-scale contexts to achieve good local reasoning. Previous works often address it as two-class segmentation problems utilizing complicated multi-step procedures, including refinement networks and complex graphical models. We argue that semantic salient segmentation can instead be effectively resolved by reformulating it as a simple yet intuitive pixel-pair-based connectivity prediction task. Following the intuition that salient objects can be naturally grouped via semantic-aware connectivity between neighboring pixels, we propose a pure Connectivity Net (ConnNet). ConnNet predicts the connectivity probabilities of each pixel with its neighboring pixels by leveraging multi-level cascade contexts embedded in the image and long-range pixel relations. We investigate our approach on two tasks, namely, salient object segmentation and salient instance-level segmentation, and illustrate that consistent improvements can be obtained by modeling these tasks as connectivity instead of binary segmentation tasks for a variety of network architectures. We achieve the state-of-the-art performance, outperforming or being comparable to existing approaches while reducing inference time due to our less complex approach.en_US
dc.descriptionAccepted manuscript version. Published version available at <a href=https://doi.org/10.1109/TIP.2018.2886997>https://doi.org/10.1109/TIP.2018.2886997.</a> © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.en_US
dc.identifier.citationKampffmeyer, M., Dong, N., Liang, X., Zhang, Y. & Xing, E.P. (2018). ConnNet: A Long-Range Relation-Aware Pixel-Connectivity Network for Salient Segmentation. <i>IEEE Transactions on Image Processing, 28</i>(5), 2518-2529. https://doi.org/10.1109/TIP.2018.2886997en_US
dc.identifier.cristinIDFRIDAID 1655951
dc.identifier.doi10.1109/TIP.2018.2886997
dc.identifier.issn1057-7149
dc.identifier.issn1941-0042
dc.identifier.urihttps://hdl.handle.net/10037/15212
dc.language.isoengen_US
dc.publisherInstitute of Electrical and Electronics Engineers (IEEE)en_US
dc.relation.journalIEEE Transactions on Image Processing
dc.relation.projectIDNorges forskningsråd: 239844en_US
dc.relation.projectIDinfo:eu-repo/grantAgreement/RCN/IKTPLUSS/239844/Norway/Next Generation Kernel-Based Machine Learning for Big Missing Data Applied to Earth Observation//en_US
dc.rights.accessRightsopenAccessen_US
dc.subjectSalient segmentationen_US
dc.subjectconvolutional neural networksen_US
dc.subjectsalient instance-level segmentationen_US
dc.subjectconnectivityen_US
dc.subjectVDP::Mathematics and natural science: 400::Physics: 430en_US
dc.subjectVDP::Matematikk og Naturvitenskap: 400::Fysikk: 430en_US
dc.titleConnNet: A Long-Range Relation-Aware Pixel-Connectivity Network for Salient Segmentationen_US
dc.typeJournal articleen_US
dc.typeTidsskriftartikkelen_US
dc.typePeer revieweden_US


File(s) in this item

Thumbnail

This item appears in the following collection(s)

Show simple item record