A Fast Alignment Scheme for Automatic OCR Evaluation of Books

Yalnız, İsmet Zeki; Manmatha, R.

doi:10.1109/icdar.2011.157

Cited by 51 publications

(39 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We have recently learned about a similar tool, RETAS [14] . This starts its alignment based on the neat idea that texts through their Zipf distribution typically have about 50% of their word types being hapaxes.…”

Section: The Problem Of Aligning 'Old' With 'Gold'mentioning

confidence: 99%

On OCR ground truths and OCR post-correction gold standards, tools and formats

Reynaert

2014

Proceedings of the First International Conference on Digital Access to Textual Cultural Heritage

View full text Add to dashboard Cite

We give an overview of activities undertaken in the sidelines of our automatic OCR post-correction core business over the past few years. We present ongoing projects in the Netherlands in which Text-Induced Corpus Clean-up plays a part. We describe the infrastructure we are building to help improve the overall text quality of large digitized text collections. We provide information on the tools we develop to facilitate the process and discuss the role of FoLiA XML which we adopted as a pivot format. Connecting the dots, we discuss the difference we perceive between OCR ground truths and OCR post-correction gold standards and their respective contributions.

show abstract

Section: The Problem Of Aligning 'Old' With 'Gold'mentioning

confidence: 99%

On OCR ground truths and OCR post-correction gold standards, tools and formats

Reynaert

2014

Proceedings of the First International Conference on Digital Access to Textual Cultural Heritage

View full text Add to dashboard Cite

show abstract

“…For evaluation purposes, a noise-free version of the same text is downloaded from the Project Gutenberg's website 2 . For labeling word bounding boxes, the OCR output and the ground truth text are aligned using a text alignment tool [15]. The estimated character accuracy for the whole book is 98.4%.…”

Section: A Datasetsmentioning

confidence: 99%

An Efficient Framework for Searching Text in Noisy Document Images

Yalnız

Manmatha

2012

2012 10th IAPR International Workshop on Document Analysis Systems

Self Cite

View full text Add to dashboard Cite

Abstract-An efficient word spotting framework is proposed to search text in scanned books. The proposed method allows one to search for words when optical character recognition (OCR) fails due to noise or for languages where there is no OCR. Given a query word image, the aim is to retrieve matching words in the book sorted by the similarity. In the offline stage, SIFT descriptors are extracted over the corner points of each word image. Those features are quantized into visual terms (visterms) using hierarchical K-Means algorithm and indexed using an inverted file. In the query resolution stage, the candidate matches are efficiently identified using the inverted index. These word images are then forwarded to the next stage where the configuration of visterms on the image plane are tested. Configuration matching is efficiently performed by projecting the visterms on the horizontal axis and searching for the Longest Common Subsequence (LCS) between the sequences of visterms. The proposed framework is tested on one English and two Telugu books. It is shown that the proposed method resolves a typical user query under 10 milliseconds providing very high retrieval accuracy (Mean Average Precision 0.93). The search accuracy for the English book is comparable to searching text in the high accuracy output of a commercial OCR engine.

show abstract

“…Feng [1] and Yalniz [8] both present different methods of aligning ground truth with OCR that do not rely on the existence of a canonically correct electronic document. Instead, both assume their ground truth texts to be in the form of plain text, typically ASCII.…”

Section: Related Workmentioning

confidence: 99%

Enhancing the Searchability of Page-Image PDF Documents Using an Aligned Hidden Layer from a Truth Text

Brailsford

2016

Proceedings of the 2016 ACM Symposium on Document Engineering

View full text Add to dashboard Cite

The search accuracy achieved in a PDF image-plus-hiddentext (PDF-IT) document depends upon the accuracy of the optical character recognition (OCR) process that produced the searchable hidden text layer. In many cases recognising words in a blurred area of a PDF page image may exceed the capabilities of an OCR engine.This paper describes a project to replace an inadequate hidden textual layer of a PDF-IT file with a more accurate hidden layer produced from a 'truth text'. The alignment of the truth text with the image is guided by using OCRprovided page-image co-ordinates, for those glyphs that are correctly recognised, as a set of fixed location points between which other truth-text words can be inserted and aligned with blurred glyphs in the image. Results are presented to show the much enhanced searchability of this new file when compared to that of the original file, which had an OCRproduced hidden layer with no truth-text enhancement.

show abstract

A Fast Alignment Scheme for Automatic OCR Evaluation of Books

Cited by 51 publications

References 5 publications

On OCR ground truths and OCR post-correction gold standards, tools and formats

On OCR ground truths and OCR post-correction gold standards, tools and formats

An Efficient Framework for Searching Text in Noisy Document Images

Enhancing the Searchability of Page-Image PDF Documents Using an Aligned Hidden Layer from a Truth Text

Contact Info

Product

Resources

About