Graph Embedded One-Class Classifiers for media data classification

作者：

Highlights：

• Introducing the Graph-Embedded One-Class SVM and Graph Embedded SVDD methods.

• Exploiting data relationships in the optimization process.

• Adopting generic geometric class information acts as a regularizer.

• The regularized solutions for both proposed methods are equivalent.

• They compare favorably to the standard OC-SVM and SVDD classifiers.

摘要

Highlights•Introducing the Graph-Embedded One-Class SVM and Graph Embedded SVDD methods.•Exploiting data relationships in the optimization process.•Adopting generic geometric class information acts as a regularizer.•The regularized solutions for both proposed methods are equivalent.•They compare favorably to the standard OC-SVM and SVDD classifiers.

论文关键词：Media data classification,One-Class Support Vector Machine,Support Vector Data Description,Graph-based regularization

论文评审过程：Received 6 November 2015, Revised 22 April 2016, Accepted 28 May 2016, Available online 3 June 2016, Version of Record 28 June 2016.

论文官网地址：https://doi.org/10.1016/j.patcog.2016.05.033