Automatic segmentation technique for acetabulum and femoral head in CT images
作者：
Highlights：
• Classify the hip joints into four groups based on anatomical and imaging criteria.
• Divide the image set into bone and non-bone classes for the initial segmentation.
• Present a Bayes decision rule for segmenting the femoral head and acetabulum.
• Refine the bone boundaries based on normal direction of vertices of bone surface.
• Evaluate our method based on clinical application criteria (good, moderate and poor).
摘要
•Classify the hip joints into four groups based on anatomical and imaging criteria.•Divide the image set into bone and non-bone classes for the initial segmentation.•Present a Bayes decision rule for segmenting the femoral head and acetabulum.•Refine the bone boundaries based on normal direction of vertices of bone surface.•Evaluate our method based on clinical application criteria (good, moderate and poor).
论文关键词：Hip joint,Osteoarthritis,Mathematical morphology,Vertex normal,Threshold selection
论文评审过程：Received 12 July 2011, Revised 30 November 2012, Accepted 6 April 2013, Available online 17 April 2013.
论文官网地址：https://doi.org/10.1016/j.patcog.2013.04.006
