Writer Identification using Deep Learning with FAST Keypoints and Harris corner detector

作者：

Highlights：

• Writer identification from handwriting samples using machine-learned features.

• Feature extraction using patches around Harris and FAST key points in handwriting.

• Encoding CNN based features using VLAD and Triangular embedding.

• Experimental study on three benchmark datasets, IAM, QUWI and IFN/ENIT.

摘要

•Writer identification from handwriting samples using machine-learned features.•Feature extraction using patches around Harris and FAST key points in handwriting.•Encoding CNN based features using VLAD and Triangular embedding.•Experimental study on three benchmark datasets, IAM, QUWI and IFN/ENIT.

论文关键词：Writer identification,Handwriting keypoints,Feature encoding,Convolutional Neural Networks

论文评审过程：Received 1 October 2020, Revised 24 March 2021, Accepted 22 June 2021, Available online 1 July 2021, Version of Record 8 July 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115473