Scaling Up Whole-Book Recognition

Xiu, Pingping; Baird, Henry S.

doi:10.1109/icdar.2009.22

Cited by 9 publications

(11 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Words in the dictionary are expanded by appending all their doublets at their end in the same order. Except for those changes, word recognition and iconic model adaptation remain exactly the same as introduced in [5].…”

Section: Methodsmentioning

confidence: 89%

“…The whole process works this way: in Stage 1, apply iconic model adaptation algorithm as in [5] until it converges; in Stage 2, alternate between adapting linguistic and iconic models multiple times.…”

Section: Methodsmentioning

confidence: 99%

“…In our previous work, algorithms that conduct only iconic model adaptation work well with a perfect (fixed) linguistic model [5] [6]. Theoretical analysis [6] has revealed that this is due in part to the effectiveness of contextual prediction, a process of predicting a character image's class from the context of its word image together with the linguistic model.…”

Section: Introductionmentioning

confidence: 97%

“…When this strategy is integrated into whole-book recognition, the whole algorithm consists of two stages: Stage 1, in which the iconic model adaptation pushes error rates down to its limit (previous research demonstrated that the iconic model adaptation algorithm yields progressively better classification results during its execution [5]); Stage 2, in which adaptations on the iconic model and the linguistic model are alternately applied multiple times.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Incorporating Linguistic Model Adaptation into Whole-Book Recognition

Xiu

Baird

2010

2010 20th International Conference on Pattern Recognition

Self Cite

View full text Add to dashboard Cite

Whole-book recognition is a document image analysis strategy that operates on the complete set of a book's page images using automatic adaptation to improve accuracy. Our algorithm expects to be given approximate iconic and linguistic models-derived from (generally errorful) OCR results and (generally incomplete) dictionaries-and then, guided entirely by evidence internal to the test set, corrects the models yielding improved accuracy. The iconic model describes image formation and determines the behavior of a character-image classifier. The linguistic model describes word-occurrence probabilities. In previous work, we reported that adapting the iconic model alone (with a perfect linguistic model) was able to automatically reduce word error rate on a 180-page book by a large factor. In this paper, we propose an algorithm that adapts both the iconic model and the linguistic model alternately to improve both models on the fly. The linguistic model adaptation method, which we report here, identifies new words and adds them to the dictionary. With 64.6% words missing in the initial dictionary, our previous algorithm reduced word error rate from 40.2% to 23.2%. The new algorithm drives word error rate down further from 23.2% to 16.0%.

show abstract

Section: Methodsmentioning

confidence: 89%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 97%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Incorporating Linguistic Model Adaptation into Whole-Book Recognition

Xiu

Baird

2010

2010 20th International Conference on Pattern Recognition

Self Cite

View full text Add to dashboard Cite

show abstract

“…Rasagna et al [27] cluster word images in an entire Telugu book using locality sensitive hashing and use this to correct character labels based on majority voting. Xiu and Baird [36], measure the disagreements between OCR results and language models using mutual entropy across a passage. This measure is then used to correct frequent OCR errors, or to add new words to the language model.…”

Section: Previous Workmentioning

confidence: 99%

Nearest neighbor based collection OCR

Sankar

Jawahar

Manmatha

2010

Proceedings of the 9th IAPR International Workshop on Document Analysis Systems

View full text Add to dashboard Cite

Conventional optical character recognition (OCR) systems operate on individual characters and words, and do not normally exploit document or collection context. We describe a Collection OCR which takes advantage of the fact that multiple examples of the same word (often in the same font) may occur in a document or collection. The idea here is that an OCR or a reCAPTCHA like process generates a partial set of recognized words. In the second stage, a nearest neighbor algorithm compares the remaining word-images to those already recognized and propagates labels from the nearest neighbors. It is shown that by using an approximate fast nearest neighbor algorithm based on Hierarchical K-Means (HKM), we can do this accurately and efficiently. It is also shown that profile based features perform much better than SIFT and Pyramid Histogram of Gradient (PHOG) features. We believe that this is because profile features are more robust to word degradations (common in our documents). This approach is applied to a collection of Telugu booksa language for which no commercial OCR exists. We show from a selection of 33 Telugu books that starting with OCR labels for only 30% of the collection we can recognize the remaining 70% of the words in the collection with 70% accuracy using this approach. Since the approach makes no language specific assumptions, it should be applicable to a large number of languages. In particular we are interested in its applicability to Indic languages and scripts.

show abstract