HEGM: A hierarchical elastic graph matching for hand gesture recognition

作者：

Highlights：

• The HEGM algorithm classifies hand postures with 99.85% recognition accuracy.

• Elastic graph nodes′ weights reflect power of inter-class discrimination.

• Using hierarchy level information enables substantial computing cycles saving.

• An innovative semi-automatic annotation technique allows fast graph creation.

摘要

•The HEGM algorithm classifies hand postures with 99.85% recognition accuracy.•Elastic graph nodes′ weights reflect power of inter-class discrimination.•Using hierarchy level information enables substantial computing cycles saving.•An innovative semi-automatic annotation technique allows fast graph creation.

论文关键词：Elastic bunch graph,Graph matching,Feature hierarchy,Feature extraction,Hand gesture recognition

论文评审过程：Available online 25 June 2013.

论文官网地址：https://doi.org/10.1016/j.patcog.2013.05.028