A flexible framework for online document segmentation by pairwise stroke distance learning

作者：

Highlights：

• A flexible framework for handling segmentation of online documents is introduced.

• The system is based on single linkage clustering with globally trained distance.

• Absence of external parameters permits flexibility for various segmentation tasks.

• Results on a variety of segmentation tasks demonstrate the method validity.

摘要

Highlights•A flexible framework for handling segmentation of online documents is introduced.•The system is based on single linkage clustering with globally trained distance.•Absence of external parameters permits flexibility for various segmentation tasks.•Results on a variety of segmentation tasks demonstrate the method validity.

论文关键词：Document analysis and recognition,Online handwriting,Segmentation,Layout analysis,Clustering,Pen computing

论文评审过程：Received 24 May 2014, Revised 17 September 2014, Accepted 16 October 2014, Available online 27 October 2014.

论文官网地址：https://doi.org/10.1016/j.patcog.2014.10.022