Online handwritten Chinese text recognition
(OHCTR) is a challenging problem as it involves a large-scale
character set, ambiguous segmentation, and variable-length input
sequences. In this paper, we exploit the outstanding capability
of path signatu
