Turkish OCR on mobile and scanned document images

Author

Karasu, Kurtulus ; Bastan, Muhammet

Author_Institution

Bigisayar Muhendisligi Bolumu, Turgut Ozal Univ., Ankara, Turkey

fYear

2015

fDate

16-19 May 2015

Firstpage

2074

Lastpage

2077

Abstract

Optical character recognition (OCR) systems have been widely used to convert documents into digital form. There are lots of both commercial and open source OCR systems available, but a benchmark on Turkish OCR is nonexistent. In this work, we first prepared two publicly available datasets for Turkish OCR, consisting of scanned document images and mobile camera captured document images. Then, we evaluated the Turkish OCR performance of three popular open source OCR systems (Tesseract, CuneiForm, GOCR) on the datasets. Tesseract outperformed the other two on both datasets.

Keywords

document image processing; mobile computing; optical character recognition; public domain software; CuneiForm; GOCR; Tesseract; Turkish OCR systems; commercial OCR systems; mobile camera captured document images; open source OCR systems; optical character recognition systems; scanned document images; Benchmark testing; Cameras; Character recognition; Histograms; Layout; Mobile communication; Optical character recognition software; Tesseract; Turkish OCR; benchmark; dataset; mobile device; scanner;

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing and Communications Applications Conference (SIU), 2015 23th

Conference_Location

Malatya

Type

conf

DOI

10.1109/SIU.2015.7130278

Filename

7130278