Segmentation of clusters by template rotation expectation maximization

详细信息    查看全文

• 作者：Carl-Magnus Svensson ; ^a ; ^{carl-magnus.svensson@leibniz-hki.de} ; Karen Grace Bondoc^b ; Georg Pohnert^b ; Marc Thilo Figge^a ; ^c ; ^{thilo.figge@leibniz-hki.de}
• 关键词：Segmentation ; Generative model ; Expectation maximization ; Template matching
• 刊名：Computer Vision and Image Understanding
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：154
• 期：Complete
• 页码：64-72
• 全文大小：1837 K
• 卷排序：154

文摘

We introduce TREM, an algorithm for resolving clusters of nearly identical objects. TREM builds on generative model incorporating a template representing the object. Algorithm applied to microscopy data of the diatom Seminavis robusta. TREM is evaluated against three alternative algorithms.