Higher-order graph cuts and medical image segmentation

    研究成果: Conference contribution

    抄録

    Energy minimization is regularly used for medical image segmentation. Higher-order energies are perhaps not as common, but are nevertheless being used increasingly often. Whereas the common first- order (pairwise) potential can directly model only the relationship between pairs of pixels, the higher-order potential can model more complex and useful relationships between more than two variables. For instance, sets of pixels, chosen according to the shape to be segmented, can be encouraged to be entirely in one segment or the other by higher-order terms. In this talk, I will describe methods for minimizing higher-order energies using graph cuts as well as some real-world examples of their applications in medical image segmentation that have been deployed in commercial medical imaging software.

    本文言語English
    ホスト出版物のタイトルLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    出版社Springer Verlag
    9555
    ISBN(印刷版)9783319302843
    出版ステータスPublished - 2016
    イベント7th Pacific-Rim Symposium on Image and Video Technology, PSIVT 2015 - Auckland, New Zealand
    継続期間: 2015 11 232015 11 27

    出版物シリーズ

    名前Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    9555
    ISSN(印刷版)03029743
    ISSN(電子版)16113349

    Other

    Other7th Pacific-Rim Symposium on Image and Video Technology, PSIVT 2015
    国/地域New Zealand
    CityAuckland
    Period15/11/2315/11/27

    ASJC Scopus subject areas

    • コンピュータ サイエンス(全般)
    • 理論的コンピュータサイエンス

    フィンガープリント

    「Higher-order graph cuts and medical image segmentation」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

    引用スタイル