Improving airway segmentation in computed tomography using leak detection with convolutional networks

详细信息    查看全文

• 作者：Jean-Paul Charbonnier ; ^a ; ^{jean-paul.charbonnier@radboudumc.nl} ; Eva M. van Rikxoort^a ; Arnaud A.A. Setio^a ; Cornelia M. Schaefer-Prokop^c ; ^b ; Bram van Ginneken^a ; Francesco Ciompi^a
• 关键词：Airway segmentation ; Chest computed tomography ; Convolutional networks
• 刊名：Medical Image Analysis
• 出版年：2017
• 出版时间：February 2017
• 年：2017
• 卷：36
• 期：Complete
• 页码：52-60
• 全文大小：3042 K
• 卷排序：36

文摘

A novel method to improve airway segmentation in thoracic CT is presented. Convolutional networks are used to detect leaks in an airway segmentation. We propose a way to largely circumvent fine-tuning of an airway segmentation algorithm. Our method was validated on an independent set of the EXACT09 challenge. We increase the tree length while limiting leaks compared to other methods in EXACT09.