Benchmarking human epithelial type 2 interphase cells classification methods on a very large dataset

作者：

Highlights：

• Benchmarking of 14 HEp-2 cell images classification methods on a very large dataset.

• Six HEp-2 staining patterns considered: four common patterns plus two challenging rare patterns.

• Best methods use features extracted from local statistics of an image.

• SVM is the most effective classifier for this problem.

摘要

Highlights•Benchmarking of 14 HEp-2 cell images classification methods on a very large dataset.•Six HEp-2 staining patterns considered: four common patterns plus two challenging rare patterns.•Best methods use features extracted from local statistics of an image.•SVM is the most effective classifier for this problem.

论文关键词：Large-scale benchmarking,Computer-aided diagnosis systems,Indirect immunofluorescence,Hep-2 cell classification

论文评审过程：Received 12 September 2014, Revised 31 July 2015, Accepted 3 August 2015, Available online 13 August 2015, Version of Record 14 November 2015.

论文官网地址：https://doi.org/10.1016/j.artmed.2015.08.001