Evaluating Performance of Nepali Script OCR using Tesseract and Artificial Neural Network

2018 IEEE 3rd International Conference on Computing, Communication and Security (ICCCS) Pub Date : 2018-10-01 DOI:10.1109/CCCS.2018.8586808

Sudan Prajapati, S. R. Joshi, Aman Maharjan, Bikash Balami

引用次数: 1

Abstract

This paper evaluates the performance of Nepali Script OCR using Tesseract and ANN. A dataset of 69 Nepali fonts with the 2,484 character samples of consonants was used in the study. With Tesseract, the overall accuracy of 96% was obtained in the training phase and 69% in the testing phase. Similarly, with ANN, an accuracy of 98% was obtained in training and 81% in testing.

查看原文本刊更多论文

用Tesseract和人工神经网络评价尼泊尔文字OCR的性能

本文利用Tesseract和ANN对尼泊尔文字OCR的性能进行了评价。研究中使用了69种尼泊尔语字体和2484个辅音字符样本的数据集。使用Tesseract，训练阶段的总体准确率为96%，测试阶段的总体准确率为69%。同样，使用人工神经网络，在训练中获得98%的准确率，在测试中获得81%的准确率。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

2018 IEEE 3rd International Conference on Computing, Communication and Security (ICCCS)

自引率

0.00%

发文量