Multi-parametric optic disc segmentation using superpixel based feature classification

作者：

Highlights：

• A new preprocessing pipeline is used for vessel removal and optic disc enhancement.

• Image regions are characterized by statistical and textural properties.

• A set of highly discriminative features is adapted.

• A comparison of supervised classifiers for optic disk localization is presented.

摘要

•A new preprocessing pipeline is used for vessel removal and optic disc enhancement.•Image regions are characterized by statistical and textural properties.•A set of highly discriminative features is adapted.•A comparison of supervised classifiers for optic disk localization is presented.

论文关键词：AdaBoostM1,Glaucoma,RusBoost,Random forest,Support vector machine

论文评审过程：Received 26 April 2018, Revised 3 December 2018, Accepted 4 December 2018, Available online 4 December 2018, Version of Record 12 December 2018.

论文官网地址：https://doi.org/10.1016/j.eswa.2018.12.008