References & Citations
Computer Science > Computer Vision and Pattern Recognition
Title: Capsule-Based Persian/Arabic Robust Handwritten Digit Recognition Using EM Routing
(Submitted on 8 Dec 2019 (v1), last revised 18 Dec 2019 (this version, v2))
Abstract: In this paper, the problem of handwritten digit recognition has been addressed. However, the underlying language is Persian/Arabic, and the system with which this task is a capsule network (CapsNet) has recently emerged as a more advanced architecture than its ancestor, namely CNN (Convolutional Neural Network). The training of the architecture is performed using the Hoda dataset, which has been provided for Persian/Arabic handwritten digits. The output of the system clearly outperforms the results achieved by its ancestors, as well as other previously presented recognition algorithms.
Submission history
From: Ali Ghofrani [view email][v1] Sun, 8 Dec 2019 07:58:26 GMT (2237kb,D)
[v2] Wed, 18 Dec 2019 20:00:22 GMT (2237kb,D)
Link back to: arXiv, form interface, contact.