JCLMM: A finite mixture model for clustering of circular-linear data and its application to psoriatic plaque segmentation
作者：
Highlights：
• Development of a novel mixture model based clustering algorithm which can deal with circular–linear bi-variate data.
• Able to deal with intra-component (with different marginal distributions) and inter-component heterogeneity among the clusters.
• The mixture model is used for color clustering in circular-linear color space to segment psoriatic plaques in skin images.
• A comparative study is presented to bring out the viability of the current approach.
摘要
Highlights•Development of a novel mixture model based clustering algorithm which can deal with circular–linear bi-variate data.•Able to deal with intra-component (with different marginal distributions) and inter-component heterogeneity among the clusters.•The mixture model is used for color clustering in circular-linear color space to segment psoriatic plaques in skin images.•A comparative study is presented to bring out the viability of the current approach.
论文关键词：Mixture model,Circular–linear data,Expectation maximization (EM),Psoriasis,Psoriatic plaque,Segmentation
论文评审过程：Received 28 October 2016, Accepted 14 December 2016, Available online 3 January 2017, Version of Record 12 March 2017.
论文官网地址：https://doi.org/10.1016/j.patcog.2016.12.016
