2020 Undergraduate Research Showcase
 

Document Type

Student Presentation

Presentation Date

4-24-2020

Faculty Sponsor

Dr. Elisa H. Barney Smith

Abstract

The idea of segmentation-free handwriting recognition has been introduced within the rise of deep learning. This technique is designed to recognize any script language/symbols as long as feedable training image set exists. The VGG-16 convolutional neural network model is used as a character spotting network using Faster R-CNN. Through the process of manual tagging, the location, size, and types of recognizable symbols are provided to train the network. This approach has been tested previously on text written in the Bangla script, where it has shown over 90% of accuracy overall. For Bangla, the network is trained and tested on Boise State Bangla Handwriting dataset. For Korean, the network is trained using the PE_92 Handwritten Korean character image database and shows promising results.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.