• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Society of Photo-Optical Instrumentation Engineers;SPIE Medical Imaging Conference >Automatic delineation of anterior and posterior cruciate ligaments by combining deep learning and deformable atlas based segmentation
【24h】

Automatic delineation of anterior and posterior cruciate ligaments by combining deep learning and deformable atlas based segmentation

机译：通过结合深度学习和基于可变形图集的分割，自动描绘前十字韧带和后十字韧带

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Quantitative evaluation of bones and ligaments around knee joint from magnetic resonance imaging (MRI) often requiresthe boundaries of selected structures to be manually traced using computer software. It may take several hours todelineate all structures of interest in a three-dimensional (3D) dataset used for the evaluation. Thus, providing automatedtools, which can delineate knee anatomical structures can improve productivity and efficiency in radiology departments.In recent years, 3D deep convolutional neural networks (3D CNN) have been successfully used for segmentation of kneebones and cartilage. However, the key challenge is segmentation of the anterior cruciate ligament (ACL) and theposterior cruciate ligament (PCL), due to high variability of intensities in the areas of pathologies such as ligament tear.In this approach, an open source 3D CNN is adapted for segmentation of knee bones and ligaments in the knee MRI. Thesegmentation accuracy of ACL and PCL is improved further by atlas based segmentation technique. The atlas mask isnon-rigidly aligned with the patient image based on composite of rigid and deformable vector field derived between thebone masks in the atlas and corresponding segmented bone masks in the patient image. The level set functionscorresponding to particular objects of interest of the deformed atlas are used to refine segmentation of the correspondingobjects in the patient image. The accuracy of the proposed method is assessed using Dice coefficient score for 50 manualsegmentations of bone, cartilage and ligaments comprising of both normal and knee injury cases. Our results show thatthe proposed approach offers a viable alternative to manual contouring of knee MRI volume by a human reader withimproved accuracy compared to the 3D CNN.
机译：通常需要通过磁共振成像（MRI）定量评估膝关节周围的骨骼和韧带 使用计算机软件手动跟踪的选定结构的边界。可能要花几个小时才能 在用于评估的三维（3D）数据集中描述所有感兴趣的结构。因此，提供自动化 可以描绘膝盖解剖结构的工具可以提高放射科的生产率和效率。 近年来，3D深度卷积神经网络（3D CNN）已成功用于膝盖的分割 骨头和软骨。然而，关键的挑战是前交叉韧带（ACL）和前交叉韧带的分割。 后交叉韧带（PCL），因为韧带撕裂等病理区域的强度变化很大。 在这种方法中，开放源3D CNN适用于在膝盖MRI中分割膝盖骨和韧带。这 基于图集的分割技术进一步提高了ACL和PCL的分割精度。地图集面具是 基于刚性和可变形矢量场之间的合成，非刚性地与患者图像对齐 图谱中的骨面具和患者图像中相应的分段骨面具。水平设定功能 与变形图集的特定感兴趣对象相对应的图像用于细化对应图像的分割 患者图像中的对象。使用50个手册的骰子系数评分来评估所提出方法的准确性 包括正常和膝关节损伤病例在内的骨骼，软骨和韧带的分割。我们的结果表明 所提出的方法为人类阅读器提供了一种可行的替代方法，代替人工绘制膝盖MRI体积的轮廓 与3D CNN相比，精度更高。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号