Handwriting posture prediction based on unsupervised model

作者：

Highlights：

• We propose a novel problem of “handwriting posture prediction”.

• We propose a neural network constructed with small convolution kernels to extract features from handwriting.

• We empirically obtain the handwriting cluster center through unsupervised learning.

• Extensive experiments reveal that our writing posture prediction approach achieves an accuracy rate of 93.3 which is significantly higher than the 76.67.

摘要

•We propose a novel problem of “handwriting posture prediction”.•We propose a neural network constructed with small convolution kernels to extract features from handwriting.•We empirically obtain the handwriting cluster center through unsupervised learning.•Extensive experiments reveal that our writing posture prediction approach achieves an accuracy rate of 93.3 which is significantly higher than the 76.67.

论文关键词：Pupils,Writing posture prediction,Features extracting,Neural network model,Unsupervised learning,Data analysis

论文评审过程：Received 28 April 2019, Revised 30 September 2019, Accepted 19 October 2019, Available online 2 November 2019, Version of Record 18 November 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.107093