An ordinal CNN approach for the assessment of neurological damage in Parkinson’s disease patients

作者：

Highlights：

• Ordinal target labels improve classification results in PD diagnosis over nominal.

• Ordinal information may be used to guide the data augmentation process.

• Data augmentation for 3D images can be performed efficiently after a CNN projection.

• The beta distribution is better suited for the OGO-SP inter-class sample generation.

摘要

•Ordinal target labels improve classification results in PD diagnosis over nominal.•Ordinal information may be used to guide the data augmentation process.•Data augmentation for 3D images can be performed efficiently after a CNN projection.•The beta distribution is better suited for the OGO-SP inter-class sample generation.

论文关键词：Artificial neural networks,Ordinal classification,Data augmentation,Computer-aided diagnosis

论文评审过程：Received 8 January 2021, Revised 26 March 2021, Accepted 21 May 2021, Available online 30 May 2021, Version of Record 10 June 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115271