OCR as a Service: An Experimental Evaluation of Google Docs OCR, Tesseract, ABBYY FineReader, and Transym

Tafti, Ahmad P.; Baghaie, Ahmadreza; Assefi, Mehdi; Arabnia, Hamid R.; Yu, Zeyun; Peissig, Peggy L.

doi:10.1007/978-3-319-50835-1_66

Cited by 59 publications

(29 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Image recognition solves the classification problem: it associates images with classes to which they belong. The recognition quality of English text is at a very high-level [9]. Since the size of hieroglyphs alphabet exceeds 10000 (while the alphabet for Latin-base languages is usually less than 100) the recognition of hieroglyphs introduces new challenges to both training and recognizing processes.…”

Section: Hieroglyphs Recognitionmentioning

confidence: 99%

See 1 more Smart Citation

Recognition of images of Korean characters using embedded networks

Ilyuhin

Sheshkus

Arlazarov

2020

Twelfth International Conference on Machine Vision (ICMV 2019)

View full text Add to dashboard Cite

Despite the significant success in the field of text recognition, complex and unsolved problems still exist in this field. In recent years, the recognition accuracy of the English language has greatly increased, while the problem of recognition of hieroglyphs has received much less attention. Hieroglyph recognition or image recognition with Korean, Japanese or Chinese characters have differences from the traditional text recognition task.This article discusses the main differences between hieroglyph languages and the Latin alphabet in the context of image recognition. A light-weight method for recognizing images of the hieroglyphs is proposed and tested on a public dataset of Korean hieroglyph images. Despite the existing solutions, the proposed method is suitable for mobile devices. Its recognition accuracy is better than the accuracy of the open-source OCR framework. The presented method of training embedded net bases on the similarities in the recognition data.

show abstract

Section: Hieroglyphs Recognitionmentioning

confidence: 99%

“…% of . To compare the results we use the recently released version of Tesseract 4.0.0 [9,25], which reaches 37.67% of . Comparing to Tesseract 4.0.0 our performance is better but still is not sufficiently good.…”

Section: Trainingmentioning

confidence: 99%

Recognition of images of Korean characters using embedded networks

Ilyuhin

Sheshkus

Arlazarov

2020

Twelfth International Conference on Machine Vision (ICMV 2019)

View full text Add to dashboard Cite

show abstract

“…For instance, previous work focused on digitization the content of documents into a machine-readable format using optical character recognition (OCR). We refer to [1], [2] for a review of current OCR systems. Existing datasets include the ICDAR Robust Reading Competitions 2011, 2013, 2015, 2017 .…”

Section: Introductionmentioning

confidence: 99%

FUNSD: A Dataset for Form Understanding in Noisy Scanned Documents

Jaume

Ekenel

Thiran

2019

2019 International Conference on Document Analysis and Recognition Workshops (ICDARW)

211

133

View full text Add to dashboard Cite

In this paper, we present a new dataset for Form Understanding in Noisy Scanned Documents (FUNSD). Form Understanding (FoUn) aims at extracting and structuring the textual content of forms. The dataset comprises 200 fully annotated real scanned forms. The documents are noisy and exhibit large variabilities in their representation making FoUn a challenging task. The proposed dataset can be used for various tasks including text detection, optical character recognition (OCR), spatial layout analysis and entity labeling/linking. To the best of our knowledge this is the first publicly available dataset with comprehensive annotations addressing the FoUn task. We also present a set of baselines and introduce metrics to evaluate performance on the FUNSD dataset. The FUNSD dataset can be downloaded at https://guillaumejaume.github. io/FUNSD/.

show abstract

“…OCR in both handwritten and typewritten texts has been extensively studied in the domain of pattern recognition and computer vision. Prominent off-the-shelf OCR tools [4] include Google Cloud OCR, 1 Tesseract, 2 ABBYY FineReader, 3 and Transym. 4 All of these aforementioned software packages, with some configuration customization, can generate sufficiently machine-readable OCR output given a high-quality input image.…”

Section: Introductionmentioning

confidence: 99%

“…Prominent off-the-shelf OCR tools [4] include Google Cloud OCR, 1 Tesseract, 2 ABBYY FineReader, 3 and Transym. 4 All of these aforementioned software packages, with some configuration customization, can generate sufficiently machine-readable OCR output given a high-quality input image. OCR results, however, can be significantly dependent on scanning noise, page layout, and image resolution [5].…”

Section: Introductionmentioning

confidence: 99%

Salience of Medical Concepts of Inside Clinical Texts and Outside Medical Records for Referred Cardiovascular Patients

Moon

Liu

Chen

et al. 2019

J Healthc Inform Res

View full text Add to dashboard Cite

Outside medical records (OMRs) accompanying referred patients are frequently sent as faxes from external healthcare providers. Accessing useful and relevant information from these OMRs in a timely manner is a challenging task due to a combination of the presence of machine-illegible information and the limited system interoperability inherent in healthcare. Little research has been done on investigating information in OMRs. This paper evaluated overlapping and non-overlapping medical concepts captured from digitally faxed OMRs for patients transferring to the Department of Cardiovascular Medicine and from clinical consultant notes generated at the Mayo Clinic. We used optical character recognition (OCR) techniques to make faxed OMRs machine-readable and used natural language processing (NLP) techniques to capture clinical concepts from both machine-readable OMRs and Mayo clinical notes. We measured the level of overlap in medical concepts between OMRs and Mayo clinical narratives in the quantitative approaches and assessed the salience of concepts specific to Cardiovascular Medicine by calculating the ratio of those mentioned concepts relative to an independent clinical corpus. Among the concepts collected from the OMRs, 11.19% of those were also present in the Mayo clinical narratives that were generated within the 3 months after their initial encounter at the Mayo Clinic. For those

show abstract

OCR as a Service: An Experimental Evaluation of Google Docs OCR, Tesseract, ABBYY FineReader, and Transym

Cited by 59 publications

References 22 publications

Recognition of images of Korean characters using embedded networks

Recognition of images of Korean characters using embedded networks

FUNSD: A Dataset for Form Understanding in Noisy Scanned Documents

Salience of Medical Concepts of Inside Clinical Texts and Outside Medical Records for Referred Cardiovascular Patients

Contact Info

Product

Resources

About