Gassman Esther E, Powell Stephanie M, Kallemeyn Nicole A, Devries Nicole A, Shivanna Kiran H, Magnotta Vincent A, Ramme Austin J, Adams Brian D, Grosland Nicole M
Department of Biomedical Engineering, Seamans Center for the Engineering Arts and Sciences, The University of Iowa, Iowa City, IA 52242, USA.
Skeletal Radiol. 2008 Apr;37(4):313-9. doi: 10.1007/s00256-007-0434-z. Epub 2008 Jan 3.
The objective was to develop tools for automating the identification of bony structures, to assess the reliability of this technique against manual raters, and to validate the resulting regions of interest against physical surface scans obtained from the same specimen.
Artificial intelligence-based algorithms have been used for image segmentation, specifically artificial neural networks (ANNs). For this study, an ANN was created and trained to identify the phalanges of the human hand.
The relative overlap between the ANN and a manual tracer was 0.87, 0.82, and 0.76, for the proximal, middle, and distal index phalanx bones respectively. Compared with the physical surface scans, the ANN-generated surface representations differed on average by 0.35 mm, 0.29 mm, and 0.40 mm for the proximal, middle, and distal phalanges respectively. Furthermore, the ANN proved to segment the structures in less than one-tenth of the time required by a manual rater.
The ANN has proven to be a reliable and valid means of segmenting the phalanx bones from CT images. Employing automated methods such as the ANN for segmentation, eliminates the likelihood of rater drift and inter-rater variability. Automated methods also decrease the amount of time and manual effort required to extract the data of interest, thereby making the feasibility of patient-specific modeling a reality.
开发用于自动识别骨结构的工具，评估该技术相对于人工评分者的可靠性，并根据从同一标本获得的物理表面扫描结果验证所得的感兴趣区域。
基于人工智能的算法已用于图像分割，特别是人工神经网络（ANN）。在本研究中，创建并训练了一个人工神经网络来识别人类手部的指骨。
人工神经网络与手动追踪器之间的相对重叠率分别为，食指近节指骨0.87、中指指骨0.82、远节指骨0.76。与物理表面扫描相比，人工神经网络生成的表面表示与近节、中节和远节指骨的平均差异分别为0.35毫米、0.29毫米和0.40毫米。此外，事实证明，人工神经网络分割结构所需的时间不到人工评分者所需时间的十分之一。
事实证明，人工神经网络是从CT图像中分割指骨的可靠且有效的方法。采用人工神经网络等自动化方法进行分割，消除了评分者偏差和评分者间变异性的可能性。自动化方法还减少了提取感兴趣数据所需的时间和人工工作量，从而使患者特异性建模成为现实。
