Characterising shape patterns using features derived from best-fitting ellipsoids
作者：
Highlights：
• A method is developed to characterise highly irregular shape patterns.
• Features are defined based on best-fitting ellipsoids.
• The method is demonstrated successfully on three biomedical data sets.
• Shape patterns are learned automatically from the data.
• Landmark points are not required.
摘要
•A method is developed to characterise highly irregular shape patterns.•Features are defined based on best-fitting ellipsoids.•The method is demonstrated successfully on three biomedical data sets.•Shape patterns are learned automatically from the data.•Landmark points are not required.
论文关键词：Shape analysis,Dimorphic yeast,Pseudohyphal growth,Cancellous bone,Marbling in beef
论文评审过程：Received 16 October 2017, Revised 15 April 2018, Accepted 10 June 2018, Available online 15 June 2018, Version of Record 22 June 2018.
论文官网地址：https://doi.org/10.1016/j.patcog.2018.06.009
