837:
825:
849:
794:
4283:
4263:
518:. This relies on the input glyph being correctly isolated from the rest of the image, and the stored glyph being in a similar font and at the same scale. This technique works best with typewritten text and does not work well when new fonts are encountered. This is the technique early physical photocell-based OCR implemented, rather directly.
433:
addition, the effectiveness of binarization influences to a significant extent the quality of character recognition, and careful decisions are made in the choice of the binarization employed for a given input image type; since the quality of the method used to obtain the binary result depends on the type of image (scanned document,
889:
rate of 5% or worse if the measurement is based on whether each whole word was recognized with no incorrect letters. Using a large enough dataset is important in a neural-network-based handwriting recognition solutions. On the other hand, producing natural datasets is very complicated and time-consuming.
33:
576:
A technique known as iterative OCR automatically crops a document into sections based on the page layout. OCR is then performed on each section individually using variable character confidence level thresholds to maximize page-level OCR accuracy. A patent from the United States Patent Office has been
173:
OCR, which could recognize text printed in virtually any font. (Kurzweil is often credited with inventing omni-font OCR, but it was in use by companies, including CompuScan, in the late 1960s and 1970s.) Kurzweil used the technology to create a reading machine for blind people to have a computer read
723:
fonts, with precisely specified sizing, spacing, and distinctive character shapes, allow a higher accuracy rate during transcription in bank check processing. Several prominent OCR engines were designed to capture text in popular fonts such as Arial or Times New Roman, and are incapable of capturing
875:
text is still not 100% accurate even where clear imaging is available. One study based on recognition of 19th- and early 20th-century newspaper pages concluded that character-by-character OCR accuracy for commercial OCR software varied from 81% to 99%; total accuracy can be achieved by human review
888:
Accuracy rates can be measured in several ways, and how they are measured can greatly affect the reported accuracy rate. For example, if word context (a lexicon of words) is not used to correct software finding non-existent words, a character error rate of 1% (99% accuracy) may result in an error
553:
use a two-pass approach to character recognition. The second pass is known as adaptive recognition and uses the letter shapes recognized with high confidence on the first pass to better recognize the remaining letters on the second pass. This is advantageous for unusual fonts or low-quality scans
926:(which is always a written-out number) is an example where using a smaller dictionary can increase recognition rates greatly. The shapes of individual cursive characters themselves simply do not contain enough information to accurately (greater than 98%) recognize all handwritten cursive script.
658:
In recent years, the major OCR technology providers began to tweak OCR systems to deal more efficiently with specific types of input. Beyond an application-specific lexicon, better performance may be had by taking into account business rules, standard expression, or rich information contained in
393:
are drawn, the direction, and the pattern of putting the pen down and lifting it. This additional information can make the process more accurate. This technology is also known as "online character recognition", "dynamic character recognition", "real-time character recognition", and "intelligent
72:
of typed, handwritten or printed text into machine-encoded text, whether from a scanned document, a photo of a document, a scene photo (for example the text on signs and billboards in a landscape photo) or from subtitle text superimposed on an image (for example: from a television broadcast).
432:
because there are two colors). The task is performed as a simple way of separating the text (or any other desired image component) from the background. The task of binarization is necessary since most commercial recognition algorithms work only on binary images, as it is simpler to do so. In
524:
decomposes glyphs into "features" like lines, closed loops, line direction, and line intersections. The extraction features reduces the dimensionality of the representation and makes the recognition process computationally efficient. These features are compared with an abstract vector-like
84:, computerized receipts, business cards, mail, printed data, or any suitable documentation – it is a common method of digitizing printed texts so that they can be electronically edited, searched, stored more compactly, displayed online, and used in machine processes such as
775:
has developed an online interface for users to correct OCRed texts in the standardized ALTO format. Crowd sourcing has also been used not to perform character recognition directly but to invite software developers to develop image processing algorithms, for example, through the use of
615: – a list of words that are allowed to occur in a document. This might be, for example, all the words in the English language, or a more technical lexicon for a specific field. This technique can be problematic if the document contains words not in the lexicon, like
918:. Higher rates of recognition of general cursive script will likely not be possible without the use of contextual or grammatical information. For example, recognizing entire words from a dictionary is easier than trying to parse individual characters from script. Reading the
221:
to extract the text from the image file captured by the device. The OCR API returns the extracted text, along with information about the location of the detected text in the original image back to the device app for further processing (such as text-to-speech) or display.
695:, that enables their interactive news team to accelerate the processing of documents that need to be reviewed. They note that it enables them to process what amounts to as many as 5,400 pages per hour in preparation for reporters to review the contents.
464:
Character isolation or segmentation – For per-character OCR, multiple characters that are connected due to image artifacts must be separated; single characters that are broken into multiple pieces due to artifacts must be
2201:
Riedl, C.; Zanibbi, R.; Hearst, M. A.; Zhu, S.; Menietti, M.; Crusan, J.; Metelsky, I.; Lakhani, K. (February 20, 2016). "Detecting
Figures and Part Labels in Patents: Competition-Based Development of Image Processing Algorithms".
115:
Early versions needed to be trained with images of each character, and worked on one font at a time. Advanced systems capable of producing a high degree of accuracy for most fonts are now common, and with support for a variety of
37:
41:
39:
35:
34:
40:
460:
Script recognition – In multilingual documents, the script may change at the level of the words and hence, identification of the script is necessary, before the right OCR can be invoked to handle the specific
864:(DOE), the Information Science Research Institute (ISRI) had the mission to foster the improvement of automated technologies for understanding machine printed documents, and it conducted the most authoritative of the
744:, which are similar to printed English characters but simplified or modified for easier recognition on the platform's computationally limited hardware. Users would need to learn how to write these special glyphs.
762:
humans to perform the character recognition can quickly process images like computer-driven OCR, but with higher accuracy for recognizing images than that obtained via computers. Practical systems include the
38:
880:
handwriting, and printed text in other scripts (especially those East Asian language characters which have many strokes for a single character) – are still the subject of active research. The
415: – if the document was not aligned properly when scanned, it may need to be tilted a few degrees clockwise or counterclockwise in order to make lines of text perfectly horizontal or vertical.
217:, OCR can be used in internet connected mobile device applications that extract text captured using the device's camera. These devices that do not have built-in OCR functionality will typically use an OCR
1500:
2109:
4157:
724:
text in these fonts that are specialized and very different from popularly used fonts. As Google
Tesseract can be trained to recognize new fonts, it can recognize OCR-A, OCR-B and MICR fonts.
487:, more sophisticated techniques are needed because whitespace between letters can sometimes be greater than that between words, and vertical lines can intersect more than one character.
2486:
1437:
640:
frequencies to correct errors, by noting that certain words are often seen together. For example, "Washington, D.C." is generally far more common in
English than "Washington DOC".
2644:
929:
Most programs allow users to set "confidence rates". This means that if the software does not achieve their desired level of accuracy, a user can be notified for manual review.
190:
was one of the first customers, and bought the program to upload legal paper and news documents onto its nascent online databases. Two years later, Kurzweil sold his company to
2804:
2611:
1176:
600:
225:
3999:
182:
and a text-to-speech synthesizer. On
January 13, 1976, the finished product was unveiled during a widely reported news conference headed by Kurzweil and the leaders of the
4332:
3405:
120:
inputs. Some systems are capable of reproducing formatted output that closely approximates the original page including images, columns, and other non-textual components.
1808:
Milyaev, Sergey; Barinova, Olga; Novikova, Tatiana; Kohli, Pushmeet; Lempitsky, Victor (2013). "Image
Binarization for End-to-End Text Understanding in Natural Images".
2390:
2272:(ſ) was often interpreted as an f, . Here's evidence of the improvements we've made since then, using the corpus operator to compare the 2009, 2012 and 2019 versions
1844:
4317:
643:
Knowledge of the grammar of the language being scanned can also help determine if a word is likely to be a verb or a noun, for example, allowing greater accuracy.
36:
2782:
1455:
982:
951:
1703:
1646:
240:
OCR engines have been developed into software applications specializing in various subjects such as receipts, invoices, checks, and legal billing documents.
2479:
2268:
When we generated the original Ngram Viewer corpora in 2009, our OCR wasn't as good . This was especially obvious in pre-19th century
English, where the
483:
is accomplished relatively simply by aligning the image to a uniform grid based on where vertical grid lines will least often intersect black areas. For
1504:
3193:
2637:
2113:
1932:
129:
626:
stream or file of characters, but more sophisticated OCR systems can preserve the original layout of the page and produce, for example, an annotated
4327:
153:
archives using an optical code recognition system. In 1931, he was granted US Patent number 1,838,389 for the invention. The patent was acquired by
3515:
3362:
1221:
1171:
2204:
907:
software, but that accuracy rate still translates to dozens of errors per page, making the technology useful only in very limited applications.
4322:
3398:
2472:
381:
OCR is generally an offline process, which analyses a static document. There are cloud based services which provide an online OCR API service.
313:
Writing instructions for vehicles by identifying CAD images in a database that are appropriate to the vehicle design as it changes in real time
1830:
1484:
138:
developed a machine that read characters and converted them into standard telegraph code. Concurrently, Edmund
Fournier d'Albe developed the
134:
Early optical character recognition may be traced to technologies involving telegraphy and creating reading devices for the blind. In 1914,
4188:
3103:
2794:
2630:
1986:
1441:
4289:
3840:
3577:
3357:
526:
4337:
2964:
2012:
1246:
1226:
899:
Web-based OCR systems for recognizing hand-printed text on the fly have become well known as commercial products in recent years (see
257:
4101:
3728:
3535:
3391:
3118:
2949:
2570:
2374:
1367:
1340:
348:
836:
824:
730:
are pre-printed boxes that encourage humans to write more legibly – one glyph per box. These are often printed in a
232:, including Latin, Cyrillic, Arabic, Hebrew, Indic, Bengali (Bangla), Devanagari, Tamil, Chinese, Japanese, and Korean characters.
1306:
457:
Line and word detection – Establishment of a baseline for word and character shapes, separating words as necessary.
4056:
2889:
183:
186:. In 1978, Kurzweil Computer Products began selling a commercial version of the optical character recognition computer program.
142:, a handheld scanner that when moved across a printed page, produced tones that corresponded to specific letters or characters.
3306:
2959:
389:. Instead of merely using the shapes of glyphs and words, this technique is able to capture motion, such as the order in which
4352:
4243:
4183:
3781:
2954:
2699:
446:
or zoning – Identification of columns, paragraphs, captions, etc. as distinct blocks. Especially important in
1287:
848:
703:
There are several techniques for solving the problem of character recognition by means other than improved OCR algorithms.
3465:
2944:
1809:
1216:
538:
534:
382:
1357:
4218:
3572:
3525:
3520:
2916:
2128:
366:
4269:
3565:
3491:
3261:
3246:
3218:
3083:
3078:
2653:
1783:
772:
1459:
892:
An example of the difficulties inherent in digitizing old text is the inability of OCR to differentiate between the "
659:
color images. This strategy is called "Application-Oriented OCR" or "Customized OCR", and has been applied to OCR of
1734:
3893:
3828:
3429:
2998:
2969:
2747:
2300:"How Good Can It Get? Analysing and Improving OCR Accuracy in Large Scale Historic Newspaper Digitisation Programs"
2110:"How to optimize results from the OCR API when extracting text from an image? - Haven OnDemand Developer Community"
1677:
377:
text, one word at a time. This is especially useful for languages where glyphs are not separated in cursive script.
4294:
4152:
3791:
3622:
3445:
2841:
2694:
1181:
861:
570:
514:
337:
Optical word recognition – targets typewritten text, one word at a time (for languages that use a
4193:
3450:
3367:
3291:
3023:
2979:
2864:
2762:
1241:
1206:
680:
443:
272:
407:
OCR software often pre-processes images to improve the chances of successful recognition. Techniques include:
2053:
4238:
4223:
3876:
3871:
3771:
3639:
3420:
3271:
3241:
2908:
2083:
1939:
1623:
1585:
Tappert, C. C.; Suen, C. Y.; Wakahara, T. (1990). "The state of the art in online handwriting recognition".
1236:
530:
386:
105:
2036:"OCR as a Service: An Experimental Evaluation of Google Docs OCR, Tesseract, ABBYY FineReader, and Transym"
4198:
3958:
3677:
3672:
3128:
2821:
2799:
2789:
2757:
2732:
764:
495:
There are two basic types of core OCR algorithm, which may produce a ranked list of candidate characters.
263:
45:
Video of the process of scanning and real-time optical character recognition (OCR) with a portable scanner
525:
representation of a character, which might reduce to one or more glyph prototypes. General techniques of
4228:
4213:
4178:
3866:
3766:
3634:
2988:
876:
or Data
Dictionary Authentication. Other areas – including recognition of hand printing,
390:
199:
175:
4096:
2364:
4248:
4203:
3649:
3594:
3440:
3435:
3341:
3017:
2993:
2846:
2547:
2537:
2512:
1878:
1718:
1661:
1392:
1261:
647:
619:. Tesseract uses its dictionary to influence the character segmentation step, for improved accuracy.
550:
546:
3823:
3801:
3550:
3545:
3503:
3455:
3321:
3251:
3208:
3164:
2936:
2926:
2921:
2809:
2187:
937:
676:
589:
508:
451:
447:
338:
101:
89:
85:
4262:
832:
database, in
English documents from 1700 to 1900, based on OCR scans for the "English 2009" corpus
747:
Zone-based OCR restricts the image to a specific part of a document. This is often referred to as
209:
environment, and in mobile applications like real-time translation of foreign-language signs on a
4208:
3786:
3615:
3331:
3203:
3068:
2831:
2814:
2672:
2384:
2231:
2213:
1836:
1602:
1256:
1251:
741:
687:
650:
algorithm has also been used in OCR post-processing to further optimize results from an OCR API.
331:
2321:
Suen, C.Y.; Plamondon, R.; Tappert, A.; Thomassen, A.; Ward, J.R.; Yamamoto, K. (May 29, 1987).
1990:
2404:
Atkinson, Kristine H. (2015). "Reinventing nonpatent literature for pharmaceutical patenting".
4274:
4066:
3718:
3589:
3582:
3336:
3048:
2856:
2767:
2560:
2421:
2370:
2327:. 3rd International Symposium on Handwriting and Computer Applications, Montreal, May 29, 1987
2322:
2136:
1904:
1826:
1480:
1363:
1336:
915:
777:
484:
285:
117:
573:
which are trained to recognize whole lines of text instead of focusing on single characters.
502:
involves comparing an image to a stored glyph on a pixel-by-pixel basis; it is also known as
4347:
4342:
4019:
4009:
3816:
3610:
3560:
3555:
3498:
3486:
3213:
3098:
3073:
2874:
2777:
2555:
2413:
2349:
Comparison of
Synthesized and Natural Datasets in Neural Network Based Handwriting Solutions
2223:
1886:
1818:
1775:
1726:
1669:
1594:
1400:
1191:
562:
480:
360:
146:
135:
2460:
Annotated bibliography of references to handwriting character recognition and pen computing
1563:
630:
that includes both the original image of the page and a searchable textual representation.
541:
are used to compare image features with stored glyph features and choose the nearest match.
4132:
4076:
3898:
3540:
3460:
3325:
3286:
3281:
3149:
2879:
2752:
2727:
2709:
1291:
1266:
1186:
903:). Accuracy rates of 80% to 90% on neat, clean hand-printed characters can be achieved by
900:
251:
206:
109:
2464:
1869:
Pati, P.B.; Ramakrishnan, A.G. (May 29, 1987). "Word Level Multi-script Identification".
2299:
1882:
1722:
1665:
1396:
807:
Please help update this article to reflect recent events or newly available information.
4106:
4071:
4061:
3886:
3644:
3470:
3033:
3013:
2737:
2284:
1329:
1310:
882:
229:
169:
started the company Kurzweil Computer Products, Inc. and continued development of omni-
93:
81:
80:
from printed paper data records – whether passport documents, invoices,
4311:
4051:
4031:
3948:
3627:
3296:
3108:
3088:
2869:
2504:
2087:
2035:
1518:
904:
759:
660:
637:
473:
297:
281:
179:
17:
2235:
1606:
4137:
3968:
3276:
2438:
2347:
1840:
1647:"Survey over image thresholding techniques and quantitative performance evaluation"
1385:
Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences
1201:
872:
731:
469:
429:
342:
326:
Optical character recognition (OCR) – targets typewritten text, one
291:
214:
166:
1501:"[javascript] Using OCR and Entity Extraction for LinkedIn Company Lookup"
1077:
1890:
1730:
1474:
4233:
4004:
3913:
3908:
3530:
3508:
3233:
3113:
2826:
2742:
2719:
2667:
2565:
1196:
616:
558:
370:
352:
97:
61:
2453:
987:
885:
is commonly used for testing systems' ability to recognize handwritten digits.
4127:
4086:
4081:
3994:
3903:
3811:
3723:
3703:
2836:
2622:
2585:
2346:
Mohseni, Maedeh Haji Agha; Azmi, Reza; Layeghi, Kamran; Maleki, Sajad (2019).
2227:
1704:"OCR binarisation and image pre-processing for searching historical documents"
1476:
Wireless Communications, Networking and Applications: Proceedings of WCNA 2014
1211:
914:
is an active area of research, with recognition rates even lower than that of
668:
623:
585:
581:
434:
412:
247:
210:
187:
77:
2140:
2013:"The basic pattern recognition and classification with openCV | Damiles"
1540:
4122:
4091:
3989:
3833:
3796:
3733:
3687:
3682:
3667:
2704:
2527:
2188:"What is the point of an online interactive OCR text editor? - Fenno-Ugrica"
1231:
1166:
768:
529:
are applicable to this type of OCR, which is commonly seen in "intelligent"
425:
418:
150:
139:
3383:
2425:
1760:
1404:
1822:
4024:
3856:
3179:
3159:
3144:
3123:
3093:
3038:
3003:
2884:
2590:
2580:
2575:
421: – removal of positive and negative spots, smoothing edges
195:
170:
307:
anti-bot systems, though these are specifically designed to prevent OCR.
205:
In the 2000s, OCR was made available online as a service (WebOCR), in a
4147:
3984:
3938:
3861:
3761:
3756:
3708:
3316:
3174:
3154:
3028:
2772:
2687:
2595:
2532:
2252:
1811:
2013 12th International Conference on Document Analysis and Recognition
957:
911:
877:
737:
672:
664:
612:
566:
374:
356:
304:
65:
2417:
2129:"How We Sped Through 900 Pages of Cohen Documents in Under 10 Minutes"
1779:
1673:
4162:
4142:
4014:
3806:
2682:
2677:
2459:
2269:
1964:
1598:
923:
893:
853:
841:
829:
691:
has adapted the OCR technology into a proprietary tool they entitle
2285:"Code and Data to evaluate OCR accuracy, originally from UNLV/ISRI"
440:
Line removal – Cleaning up non-glyph boxes and lines
424:
Binarization – conversion of an image from color or
3963:
3943:
3933:
3928:
3923:
3918:
3881:
3713:
3372:
3008:
2522:
2218:
972:
968:
847:
835:
823:
716:
712:
327:
191:
69:
2161:
269:
Automatically extracting key information from insurance documents
3953:
2894:
2517:
1702:
Gupta, Maya R.; Jacobson, Nathaniel P.; Garcia, Eric K. (2007).
1383:
d'Albe, E. E. F. (July 1, 1914). "On a Type-Reading Optophone".
1074:
964:
720:
611:
OCR accuracy can be increased if the output is constrained by a
593:
3387:
2626:
2468:
2363:
Kapidakis, Sarantos; Mazurek, Cezary and Werla, Marcin (2015).
290:
Making electronic images of printed documents searchable, e.g.
149:
developed what he called a "Statistical Machine" for searching
3169:
1768:
IEEE Transactions on Pattern Analysis and Machine Intelligence
1587:
IEEE Transactions on Pattern Analysis and Machine Intelligence
787:
627:
316:
Making scanned documents searchable by converting them to PDFs
218:
154:
1359:
Advanced Image-Based Spam Detection and Filtering Techniques
1068:
963:
Some of these characters are mapped from fonts specific to
296:
Converting handwriting in real-time to control a computer (
2324:
Future Challenges in Handwriting and Computer Applications
2205:
International Journal on Document Analysis and Recognition
932:
An error introduced by OCR scanning is sometimes termed a
844:
database, based on OCR scans for the "English 2012" corpus
599:
For a list of optical character recognition software, see
310:
Assistive technology for blind and visually impaired users
359:
text one glyph or character at a time, usually involving
960:
Standard in June 1993, with the release of version 1.1.
277:
Extracting business card information into a contact list
557:
As of December 2016, modern OCR software includes
2366:
Research and Advanced Technology for Digital Libraries
2054:"How the Best OCR Technology Captures 99.91% of Data"
1926:
1924:
1922:
554:
where the font is distorted (e.g. blurred or faded).
280:
Creating textual versions of printed documents, e.g.
2612:
Comparison of optical character recognition software
1624:"Optical Character Recognition (OCR) – How it works"
1177:
Comparison of optical character recognition software
601:
Comparison of optical character recognition software
4171:
4115:
4044:
3977:
3849:
3749:
3742:
3696:
3660:
3603:
3479:
3419:
3350:
3305:
3260:
3232:
3192:
3137:
3059:
3047:
2978:
2935:
2907:
2855:
2718:
2660:
2604:
2546:
2503:
254:, passports, invoices, bank statements and receipts
1761:"Goal-directed evaluation of binarisation methods"
1438:"Extracting text from images using OCR on Android"
1328:
1080:
2454:Unicode OCR – Hex Range: 2440-245F
1331:The history of OCR, optical character recognition
1071:
580:The OCR result can be stored in the standardized
2439:http://www.hoopoes.com/jargon/entry/scanno.shtml
1564:"John Resig – OCR and Neural Nets in JavaScript"
1335:. : Recognition Technologies Users Association.
856:in English 2012 or later are normalized to an S.
369:(IWR) – also targets handwritten
351:(ICR) – also targets handwritten
1059:
734:which can be easily removed by the OCR system.
226:Various commercial and open source OCR systems
3399:
2638:
2480:
1062:
952:Optical Character Recognition (Unicode block)
174:text to them out loud. The device included a
8:
2389:: CS1 maint: multiple names: authors list (
1618:
1616:
1151:Grey areas indicate non-assigned code points
956:Characters to support OCR were added to the
437:image, degraded historical document, etc.).
3746:
3406:
3392:
3384:
3056:
2852:
2645:
2631:
2623:
2487:
2473:
2465:
2078:
2076:
2074:
1759:Trier, Oeivind Due; Jain, Anil K. (1995).
1322:
1320:
977:
4333:Automatic identification and data capture
2217:
1933:"An Overview of the Tesseract OCR Engine"
1065:
130:Timeline of optical character recognition
2456:Optical Character Recognition in Unicode
1456:"[Tutorial] OCR on Google Glass"
840:Occurrence of laft and last in Google's
828:Occurrence of laft and last in Google's
31:
1645:Sezgin, Mehmet; Sankur, Bulent (2004).
1362:. Hershey, PA: IGI Global. p. 91.
1278:
1222:Live ink character recognition solution
1172:Applications of artificial intelligence
1053:
740:used a special set of glyphs, known as
303:Defeating or testing the robustness of
4318:Optical character recognition software
2382:
2088:"How does OCR document scanning work?"
1850:from the original on November 13, 2017
1056:
1050:
988:Official Unicode Consortium code chart
145:In the late 1920s and into the 1930s,
2247:
2245:
2015:. Blog.damiles.com. November 14, 2008
1907:. Blog.damiles.com. November 20, 2008
1789:from the original on October 16, 2015
213:. With the advent of smartphones and
7:
4244:Generative adversarial network (GAN)
3104:Simple Knowledge Organization System
1905:"Basic OCR in OpenCV | Damiles"
527:feature detection in computer vision
27:Computer recognition of visual text
1458:. October 23, 2014. Archived from
1247:Outline of artificial intelligence
1227:Magnetic ink character recognition
654:Application-specific optimizations
258:Automatic number-plate recognition
194:, which eventually spun it off as
25:
3119:Thesaurus (information retrieval)
2571:Microsoft Office Document Imaging
349:Intelligent character recognition
4282:
4281:
4261:
1562:Resig, John (January 23, 2009).
1356:Dhavale, Sunita Vikrant (2017).
792:
588:maintained by the United States
184:National Federation of the Blind
100:. OCR is a field of research in
4328:Applications of computer vision
2034:Assefi, Mehdi (December 2016).
1503:. July 22, 2014. Archived from
1440:. June 27, 2015. Archived from
592:. Other common formats include
4194:Recurrent neural network (RNN)
4184:Differentiable neural computer
2700:Natural language understanding
1543:. Cs.sfu.ca. December 10, 2002
1128:
1126:
1124:
1122:
1120:
1118:
1116:
1114:
1112:
1110:
1108:
1106:
1104:
1102:
1100:
1098:
1091:
1089:
1087:
1085:
1083:
533:and most modern OCR software.
243:The software can be used for:
228:are available for most common
1:
4323:Optical character recognition
4239:Variational autoencoder (VAE)
4199:Long short-term memory (LSTM)
3466:Computational learning theory
3224:Optical character recognition
2496:Optical character recognition
2406:Pharmaceutical Patent Analyst
2127:Fehr, Tiff (March 26, 2019).
1654:Journal of Electronic Imaging
1217:List of emerging technologies
983:Optical Character Recognition
539:k-nearest neighbors algorithm
535:Nearest neighbour classifiers
428:to black-and-white (called a
383:Handwriting movement analysis
345:). Usually just called "OCR".
250:for business documents, e.g.
50:Optical character recognition
4219:Convolutional neural network
2917:Multi-document summarization
2352:. ITCT – via Civilica.
1891:10.1016/j.patrec.2008.01.027
1731:10.1016/j.patcog.2006.04.043
1521:. andrewt.net. June 28, 2006
1327:Schantz, Herbert F. (1982).
871:Recognition of typewritten,
367:Intelligent word recognition
4214:Multilayer perceptron (MLP)
3247:Latent Dirichlet allocation
3219:Natural language generation
3084:Machine-readable dictionary
3079:Linguistic Linked Open Data
2654:Natural language processing
2298:Holley, Rose (April 2009).
2253:"Google Books Ngram Viewer"
1989:. OCRWizard. Archived from
1871:Pattern Recognition Letters
1541:"Breaking a Visual CAPTCHA"
866:Annual Test of OCR Accuracy
852:Searching for words with a
773:National Library of Finland
622:The output stream may be a
565:, and Transym. Others like
4369:
4290:Artificial neural networks
4204:Gated recurrent unit (GRU)
3430:Differentiable programming
2999:Explicit semantic analysis
2748:Deep linguistic processing
1143:As of Unicode version 16.0
1132:
980:
949:
936:(by analogy with the term
127:
4338:Computational linguistics
4257:
3623:Artificial neural network
3446:Automatic differentiation
2842:Word-sense disambiguation
2695:Computational linguistics
2369:. Springer. p. 257.
2228:10.1007/s10032-016-0260-8
1182:Computational linguistics
862:U.S. Department of Energy
801:This article needs to be
262:Passport recognition and
76:Widely used as a form of
3451:Neuromorphic engineering
3414:Differentiable computing
3368:Natural Language Toolkit
3292:Pronunciation assessment
3194:Automatic identification
3024:Latent semantic analysis
2980:Distributional semantics
2865:Compound-term processing
2763:Named-entity recognition
1987:"How OCR Software Works"
1420:Data Processing Magazine
1242:Optical mark recognition
1207:Institutional repository
681:automobile manufacturing
577:issued for this method.
394:character recognition".
385:can be used as input to
273:Traffic-sign recognition
54:optical character reader
4224:Residual neural network
3640:Artificial Intelligence
3272:Automated essay scoring
3242:Document classification
2909:Automatic summarization
1519:"How To Crack Captchas"
1237:OCR in Indian Languages
531:handwriting recognition
387:handwriting recognition
161:Visually impaired users
106:artificial intelligence
3129:Universal Dependencies
2822:Terminology extraction
2805:Semantic decomposition
2800:Semantic role labeling
2790:Part-of-speech tagging
2758:Information extraction
2743:Coreference resolution
2733:Collocation extraction
2287:. Google Code Archive.
2162:"Train Your Tesseract"
1473:Zeng, Qing-An (2015).
1418:"The History of OCR".
1405:10.1098/rspa.1914.0061
896:" and "f" characters.
857:
845:
833:
778:rank-order tournaments
765:Amazon Mechanical Turk
634:Near-neighbor analysis
264:information extraction
46:
4353:Machine learning task
4179:Neural Turing machine
3767:Human image synthesis
2890:Sentence segmentation
1945:on September 28, 2010
1823:10.1109/ICDAR.2013.33
1305:OnDemand, HPE Haven.
1286:OnDemand, HPE Haven.
851:
839:
827:
200:Nuance Communications
44:
18:Character recognition
4270:Computer programming
4249:Graph neural network
3824:Text-to-video models
3802:Text-to-image models
3650:Large language model
3635:Scientific computing
3441:Statistical manifold
3436:Information geometry
3342:Voice user interface
3053:datasets and corpora
2994:Document-term matrix
2847:Word-sense induction
2548:Proprietary software
2190:. February 21, 2014.
2168:. September 20, 2018
2166:Train Your Tesseract
2090:. Explain that Stuff
2086:(January 30, 2012).
1817:. pp. 128–132.
1262:Tesseract OCR engine
860:Commissioned by the
707:Forcing better input
648:Levenshtein Distance
584:format, a dedicated
448:multi-column layouts
198:, which merged with
3616:In-context learning
3456:Pattern recognition
3322:Interactive fiction
3252:Pachinko allocation
3209:Speech segmentation
3165:Google Ngram Viewer
2937:Machine translation
2927:Text simplification
2922:Sentence extraction
2810:Semantic similarity
1931:Smith, Ray (2007).
1883:2008PaReL..29.1218P
1740:on October 16, 2015
1723:2007PatRe..40..389G
1711:Pattern Recognition
1683:on October 16, 2015
1666:2004JEI....13..146S
1397:1914RSPSA..90..373D
868:from 1992 to 1996.
711:Special fonts like
590:Library of Congress
509:pattern recognition
102:pattern recognition
90:machine translation
86:cognitive computing
4209:Echo state network
4097:Jürgen Schmidhuber
3792:Facial recognition
3787:Speech recognition
3697:Software libraries
3332:Question answering
3204:Speech recognition
3069:Corpus linguistics
3049:Language resources
2832:Textual entailment
2815:Sentiment analysis
2270:elongated medial-s
2133:The New York Times
2116:on March 22, 2016.
1993:on August 16, 2009
1965:"OCR Introduction"
1507:on April 17, 2016.
1444:on March 15, 2016.
1313:on April 19, 2016.
1294:on April 15, 2016.
1257:Speech recognition
1252:Sketch recognition
858:
846:
834:
688:The New York Times
569:and Tesseract use
522:Feature extraction
485:proportional fonts
47:
4305:
4304:
4067:Stephen Grossberg
4040:
4039:
3381:
3380:
3337:Virtual assistant
3262:Computer-assisted
3188:
3187:
2945:Computer-assisted
2903:
2902:
2895:Word segmentation
2857:Text segmentation
2795:Semantic analysis
2783:Syntactic parsing
2768:Ontology learning
2620:
2619:
2561:Adobe Acrobat Pro
2418:10.4155/ppa.15.21
1832:978-0-7695-4999-6
1780:10.1109/34.476511
1774:(12): 1191–1201.
1674:10.1117/1.1631315
1486:978-81-322-2580-5
1462:on March 5, 2016.
1157:
1156:
916:hand-printed text
901:Tablet PC history
822:
821:
677:driver's licenses
545:Software such as
515:image correlation
481:fixed-pitch fonts
468:Normalization of
286:Project Gutenberg
118:image file format
42:
16:(Redirected from
4360:
4295:Machine learning
4285:
4284:
4265:
4020:Action selection
4010:Self-driving car
3817:Stable Diffusion
3782:Speech synthesis
3747:
3611:Machine learning
3487:Gradient descent
3408:
3401:
3394:
3385:
3358:Formal semantics
3307:Natural language
3214:Speech synthesis
3196:and data capture
3099:Semantic network
3074:Lexical resource
3057:
2875:Lexical analysis
2853:
2778:Semantic parsing
2647:
2640:
2633:
2624:
2556:ABBYY FineReader
2489:
2482:
2475:
2466:
2442:
2436:
2430:
2429:
2401:
2395:
2394:
2388:
2380:
2360:
2354:
2353:
2343:
2337:
2336:
2334:
2332:
2318:
2312:
2311:
2309:
2307:
2302:. D-Lib Magazine
2295:
2289:
2288:
2281:
2275:
2274:
2265:
2263:
2257:books.google.com
2249:
2240:
2239:
2221:
2198:
2192:
2191:
2184:
2178:
2177:
2175:
2173:
2158:
2152:
2151:
2149:
2147:
2124:
2118:
2117:
2112:. Archived from
2106:
2100:
2099:
2097:
2095:
2080:
2069:
2068:
2066:
2064:
2050:
2044:
2043:
2031:
2025:
2024:
2022:
2020:
2009:
2003:
2002:
2000:
1998:
1983:
1977:
1976:
1974:
1972:
1961:
1955:
1954:
1952:
1950:
1944:
1938:. Archived from
1937:
1928:
1917:
1916:
1914:
1912:
1901:
1895:
1894:
1877:(9): 1218–1229.
1866:
1860:
1859:
1857:
1855:
1849:
1816:
1805:
1799:
1798:
1796:
1794:
1788:
1765:
1756:
1750:
1749:
1747:
1745:
1739:
1733:. Archived from
1708:
1699:
1693:
1692:
1690:
1688:
1682:
1676:. Archived from
1651:
1642:
1636:
1635:
1633:
1631:
1620:
1611:
1610:
1599:10.1109/34.57669
1582:
1576:
1575:
1573:
1571:
1559:
1553:
1552:
1550:
1548:
1537:
1531:
1530:
1528:
1526:
1515:
1509:
1508:
1497:
1491:
1490:
1470:
1464:
1463:
1452:
1446:
1445:
1434:
1428:
1427:
1415:
1409:
1408:
1391:(619): 373–375.
1380:
1374:
1373:
1353:
1347:
1346:
1334:
1324:
1315:
1314:
1309:. Archived from
1302:
1296:
1295:
1290:. Archived from
1283:
1192:Digital mailroom
1150:
1142:
978:
817:
814:
808:
796:
795:
788:
636:can make use of
563:ABBYY FineReader
504:pattern matching
491:Text recognition
479:Segmentation of
361:machine learning
147:Emanuel Goldberg
136:Emanuel Goldberg
43:
21:
4368:
4367:
4363:
4362:
4361:
4359:
4358:
4357:
4308:
4307:
4306:
4301:
4253:
4167:
4133:Google DeepMind
4111:
4077:Geoffrey Hinton
4036:
3973:
3899:Project Debater
3845:
3743:Implementations
3738:
3692:
3656:
3599:
3541:Backpropagation
3475:
3461:Tensor calculus
3415:
3412:
3382:
3377:
3346:
3326:Syntax guessing
3308:
3301:
3287:Predictive text
3282:Grammar checker
3263:
3256:
3228:
3195:
3184:
3150:Bank of English
3133:
3061:
3052:
3043:
2974:
2931:
2899:
2851:
2753:Distant reading
2728:Argument mining
2714:
2710:Text processing
2656:
2651:
2621:
2616:
2600:
2542:
2499:
2493:
2450:
2445:
2437:
2433:
2403:
2402:
2398:
2381:
2377:
2362:
2361:
2357:
2345:
2344:
2340:
2330:
2328:
2320:
2319:
2315:
2305:
2303:
2297:
2296:
2292:
2283:
2282:
2278:
2261:
2259:
2251:
2250:
2243:
2200:
2199:
2195:
2186:
2185:
2181:
2171:
2169:
2160:
2159:
2155:
2145:
2143:
2126:
2125:
2121:
2108:
2107:
2103:
2093:
2091:
2084:Woodford, Chris
2082:
2081:
2072:
2062:
2060:
2052:
2051:
2047:
2033:
2032:
2028:
2018:
2016:
2011:
2010:
2006:
1996:
1994:
1985:
1984:
1980:
1970:
1968:
1963:
1962:
1958:
1948:
1946:
1942:
1935:
1930:
1929:
1920:
1910:
1908:
1903:
1902:
1898:
1868:
1867:
1863:
1853:
1851:
1847:
1833:
1814:
1807:
1806:
1802:
1792:
1790:
1786:
1763:
1758:
1757:
1753:
1743:
1741:
1737:
1706:
1701:
1700:
1696:
1686:
1684:
1680:
1649:
1644:
1643:
1639:
1629:
1627:
1626:. Nicomsoft.com
1622:
1621:
1614:
1584:
1583:
1579:
1569:
1567:
1561:
1560:
1556:
1546:
1544:
1539:
1538:
1534:
1524:
1522:
1517:
1516:
1512:
1499:
1498:
1494:
1487:
1472:
1471:
1467:
1454:
1453:
1449:
1436:
1435:
1431:
1417:
1416:
1412:
1382:
1381:
1377:
1370:
1355:
1354:
1350:
1343:
1326:
1325:
1318:
1304:
1303:
1299:
1285:
1284:
1280:
1276:
1271:
1267:Voice recording
1187:Digital library
1162:
1147:
1139:
986:
954:
948:
910:Recognition of
818:
812:
809:
806:
797:
793:
786:
757:
709:
701:
693:Document Helper
656:
609:
607:Post-processing
571:neural networks
500:Matrix matching
493:
444:Layout analysis
405:
400:
323:
238:
230:writing systems
207:cloud computing
180:flatbed scanner
163:
132:
126:
110:computer vision
96:, key data and
82:bank statements
32:
28:
23:
22:
15:
12:
11:
5:
4366:
4364:
4356:
4355:
4350:
4345:
4340:
4335:
4330:
4325:
4320:
4310:
4309:
4303:
4302:
4300:
4299:
4298:
4297:
4292:
4279:
4278:
4277:
4272:
4258:
4255:
4254:
4252:
4251:
4246:
4241:
4236:
4231:
4226:
4221:
4216:
4211:
4206:
4201:
4196:
4191:
4186:
4181:
4175:
4173:
4169:
4168:
4166:
4165:
4160:
4155:
4150:
4145:
4140:
4135:
4130:
4125:
4119:
4117:
4113:
4112:
4110:
4109:
4107:Ilya Sutskever
4104:
4099:
4094:
4089:
4084:
4079:
4074:
4072:Demis Hassabis
4069:
4064:
4062:Ian Goodfellow
4059:
4054:
4048:
4046:
4042:
4041:
4038:
4037:
4035:
4034:
4029:
4028:
4027:
4017:
4012:
4007:
4002:
3997:
3992:
3987:
3981:
3979:
3975:
3974:
3972:
3971:
3966:
3961:
3956:
3951:
3946:
3941:
3936:
3931:
3926:
3921:
3916:
3911:
3906:
3901:
3896:
3891:
3890:
3889:
3879:
3874:
3869:
3864:
3859:
3853:
3851:
3847:
3846:
3844:
3843:
3838:
3837:
3836:
3831:
3821:
3820:
3819:
3814:
3809:
3799:
3794:
3789:
3784:
3779:
3774:
3769:
3764:
3759:
3753:
3751:
3744:
3740:
3739:
3737:
3736:
3731:
3726:
3721:
3716:
3711:
3706:
3700:
3698:
3694:
3693:
3691:
3690:
3685:
3680:
3675:
3670:
3664:
3662:
3658:
3657:
3655:
3654:
3653:
3652:
3645:Language model
3642:
3637:
3632:
3631:
3630:
3620:
3619:
3618:
3607:
3605:
3601:
3600:
3598:
3597:
3595:Autoregression
3592:
3587:
3586:
3585:
3575:
3573:Regularization
3570:
3569:
3568:
3563:
3558:
3548:
3543:
3538:
3536:Loss functions
3533:
3528:
3523:
3518:
3513:
3512:
3511:
3501:
3496:
3495:
3494:
3483:
3481:
3477:
3476:
3474:
3473:
3471:Inductive bias
3468:
3463:
3458:
3453:
3448:
3443:
3438:
3433:
3425:
3423:
3417:
3416:
3413:
3411:
3410:
3403:
3396:
3388:
3379:
3378:
3376:
3375:
3370:
3365:
3360:
3354:
3352:
3348:
3347:
3345:
3344:
3339:
3334:
3329:
3319:
3313:
3311:
3309:user interface
3303:
3302:
3300:
3299:
3294:
3289:
3284:
3279:
3274:
3268:
3266:
3258:
3257:
3255:
3254:
3249:
3244:
3238:
3236:
3230:
3229:
3227:
3226:
3221:
3216:
3211:
3206:
3200:
3198:
3190:
3189:
3186:
3185:
3183:
3182:
3177:
3172:
3167:
3162:
3157:
3152:
3147:
3141:
3139:
3135:
3134:
3132:
3131:
3126:
3121:
3116:
3111:
3106:
3101:
3096:
3091:
3086:
3081:
3076:
3071:
3065:
3063:
3054:
3045:
3044:
3042:
3041:
3036:
3034:Word embedding
3031:
3026:
3021:
3014:Language model
3011:
3006:
3001:
2996:
2991:
2985:
2983:
2976:
2975:
2973:
2972:
2967:
2965:Transfer-based
2962:
2957:
2952:
2947:
2941:
2939:
2933:
2932:
2930:
2929:
2924:
2919:
2913:
2911:
2905:
2904:
2901:
2900:
2898:
2897:
2892:
2887:
2882:
2877:
2872:
2867:
2861:
2859:
2850:
2849:
2844:
2839:
2834:
2829:
2824:
2818:
2817:
2812:
2807:
2802:
2797:
2792:
2787:
2786:
2785:
2780:
2770:
2765:
2760:
2755:
2750:
2745:
2740:
2738:Concept mining
2735:
2730:
2724:
2722:
2716:
2715:
2713:
2712:
2707:
2702:
2697:
2692:
2691:
2690:
2685:
2675:
2670:
2664:
2662:
2658:
2657:
2652:
2650:
2649:
2642:
2635:
2627:
2618:
2617:
2615:
2614:
2608:
2606:
2602:
2601:
2599:
2598:
2593:
2588:
2583:
2578:
2573:
2568:
2563:
2558:
2552:
2550:
2544:
2543:
2541:
2540:
2535:
2530:
2525:
2520:
2515:
2509:
2507:
2501:
2500:
2494:
2492:
2491:
2484:
2477:
2469:
2463:
2462:
2457:
2449:
2448:External links
2446:
2444:
2443:
2431:
2412:(5): 371–375.
2396:
2375:
2355:
2338:
2313:
2290:
2276:
2241:
2193:
2179:
2153:
2119:
2101:
2070:
2045:
2026:
2004:
1978:
1956:
1918:
1896:
1861:
1831:
1800:
1751:
1694:
1637:
1612:
1577:
1554:
1532:
1510:
1492:
1485:
1465:
1447:
1429:
1410:
1375:
1368:
1348:
1341:
1316:
1297:
1288:"OCR Document"
1277:
1275:
1272:
1270:
1269:
1264:
1259:
1254:
1249:
1244:
1239:
1234:
1229:
1224:
1219:
1214:
1209:
1204:
1199:
1194:
1189:
1184:
1179:
1174:
1169:
1163:
1161:
1158:
1155:
1154:
1153:
1152:
1144:
1130:
1129:
1127:
1125:
1123:
1121:
1119:
1117:
1115:
1113:
1111:
1109:
1107:
1105:
1103:
1101:
1099:
1097:
1093:
1092:
1090:
1088:
1086:
1084:
1082:
1079:
1076:
1073:
1070:
1067:
1064:
1061:
1058:
1055:
1052:
1049:
1045:
1044:
1041:
1038:
1035:
1032:
1029:
1026:
1023:
1020:
1017:
1014:
1011:
1008:
1005:
1002:
999:
996:
992:
991:
950:Main article:
947:
944:
883:MNIST database
820:
819:
800:
798:
791:
785:
782:
756:
753:
708:
705:
700:
697:
661:license plates
655:
652:
608:
605:
596:and PAGE XML.
543:
542:
519:
492:
489:
477:
476:
466:
462:
458:
455:
441:
438:
422:
416:
404:
403:Pre-processing
401:
399:
396:
379:
378:
364:
346:
335:
322:
319:
318:
317:
314:
311:
308:
301:
294:
288:
278:
275:
270:
267:
260:
255:
237:
234:
162:
159:
125:
122:
94:text-to-speech
92:, (extracted)
68:conversion of
26:
24:
14:
13:
10:
9:
6:
4:
3:
2:
4365:
4354:
4351:
4349:
4346:
4344:
4341:
4339:
4336:
4334:
4331:
4329:
4326:
4324:
4321:
4319:
4316:
4315:
4313:
4296:
4293:
4291:
4288:
4287:
4280:
4276:
4273:
4271:
4268:
4267:
4264:
4260:
4259:
4256:
4250:
4247:
4245:
4242:
4240:
4237:
4235:
4232:
4230:
4227:
4225:
4222:
4220:
4217:
4215:
4212:
4210:
4207:
4205:
4202:
4200:
4197:
4195:
4192:
4190:
4187:
4185:
4182:
4180:
4177:
4176:
4174:
4172:Architectures
4170:
4164:
4161:
4159:
4156:
4154:
4151:
4149:
4146:
4144:
4141:
4139:
4136:
4134:
4131:
4129:
4126:
4124:
4121:
4120:
4118:
4116:Organizations
4114:
4108:
4105:
4103:
4100:
4098:
4095:
4093:
4090:
4088:
4085:
4083:
4080:
4078:
4075:
4073:
4070:
4068:
4065:
4063:
4060:
4058:
4055:
4053:
4052:Yoshua Bengio
4050:
4049:
4047:
4043:
4033:
4032:Robot control
4030:
4026:
4023:
4022:
4021:
4018:
4016:
4013:
4011:
4008:
4006:
4003:
4001:
3998:
3996:
3993:
3991:
3988:
3986:
3983:
3982:
3980:
3976:
3970:
3967:
3965:
3962:
3960:
3957:
3955:
3952:
3950:
3949:Chinchilla AI
3947:
3945:
3942:
3940:
3937:
3935:
3932:
3930:
3927:
3925:
3922:
3920:
3917:
3915:
3912:
3910:
3907:
3905:
3902:
3900:
3897:
3895:
3892:
3888:
3885:
3884:
3883:
3880:
3878:
3875:
3873:
3870:
3868:
3865:
3863:
3860:
3858:
3855:
3854:
3852:
3848:
3842:
3839:
3835:
3832:
3830:
3827:
3826:
3825:
3822:
3818:
3815:
3813:
3810:
3808:
3805:
3804:
3803:
3800:
3798:
3795:
3793:
3790:
3788:
3785:
3783:
3780:
3778:
3775:
3773:
3770:
3768:
3765:
3763:
3760:
3758:
3755:
3754:
3752:
3748:
3745:
3741:
3735:
3732:
3730:
3727:
3725:
3722:
3720:
3717:
3715:
3712:
3710:
3707:
3705:
3702:
3701:
3699:
3695:
3689:
3686:
3684:
3681:
3679:
3676:
3674:
3671:
3669:
3666:
3665:
3663:
3659:
3651:
3648:
3647:
3646:
3643:
3641:
3638:
3636:
3633:
3629:
3628:Deep learning
3626:
3625:
3624:
3621:
3617:
3614:
3613:
3612:
3609:
3608:
3606:
3602:
3596:
3593:
3591:
3588:
3584:
3581:
3580:
3579:
3576:
3574:
3571:
3567:
3564:
3562:
3559:
3557:
3554:
3553:
3552:
3549:
3547:
3544:
3542:
3539:
3537:
3534:
3532:
3529:
3527:
3524:
3522:
3519:
3517:
3516:Hallucination
3514:
3510:
3507:
3506:
3505:
3502:
3500:
3497:
3493:
3490:
3489:
3488:
3485:
3484:
3482:
3478:
3472:
3469:
3467:
3464:
3462:
3459:
3457:
3454:
3452:
3449:
3447:
3444:
3442:
3439:
3437:
3434:
3432:
3431:
3427:
3426:
3424:
3422:
3418:
3409:
3404:
3402:
3397:
3395:
3390:
3389:
3386:
3374:
3371:
3369:
3366:
3364:
3363:Hallucination
3361:
3359:
3356:
3355:
3353:
3349:
3343:
3340:
3338:
3335:
3333:
3330:
3327:
3323:
3320:
3318:
3315:
3314:
3312:
3310:
3304:
3298:
3297:Spell checker
3295:
3293:
3290:
3288:
3285:
3283:
3280:
3278:
3275:
3273:
3270:
3269:
3267:
3265:
3259:
3253:
3250:
3248:
3245:
3243:
3240:
3239:
3237:
3235:
3231:
3225:
3222:
3220:
3217:
3215:
3212:
3210:
3207:
3205:
3202:
3201:
3199:
3197:
3191:
3181:
3178:
3176:
3173:
3171:
3168:
3166:
3163:
3161:
3158:
3156:
3153:
3151:
3148:
3146:
3143:
3142:
3140:
3136:
3130:
3127:
3125:
3122:
3120:
3117:
3115:
3112:
3110:
3109:Speech corpus
3107:
3105:
3102:
3100:
3097:
3095:
3092:
3090:
3089:Parallel text
3087:
3085:
3082:
3080:
3077:
3075:
3072:
3070:
3067:
3066:
3064:
3058:
3055:
3050:
3046:
3040:
3037:
3035:
3032:
3030:
3027:
3025:
3022:
3019:
3015:
3012:
3010:
3007:
3005:
3002:
3000:
2997:
2995:
2992:
2990:
2987:
2986:
2984:
2981:
2977:
2971:
2968:
2966:
2963:
2961:
2958:
2956:
2953:
2951:
2950:Example-based
2948:
2946:
2943:
2942:
2940:
2938:
2934:
2928:
2925:
2923:
2920:
2918:
2915:
2914:
2912:
2910:
2906:
2896:
2893:
2891:
2888:
2886:
2883:
2881:
2880:Text chunking
2878:
2876:
2873:
2871:
2870:Lemmatisation
2868:
2866:
2863:
2862:
2860:
2858:
2854:
2848:
2845:
2843:
2840:
2838:
2835:
2833:
2830:
2828:
2825:
2823:
2820:
2819:
2816:
2813:
2811:
2808:
2806:
2803:
2801:
2798:
2796:
2793:
2791:
2788:
2784:
2781:
2779:
2776:
2775:
2774:
2771:
2769:
2766:
2764:
2761:
2759:
2756:
2754:
2751:
2749:
2746:
2744:
2741:
2739:
2736:
2734:
2731:
2729:
2726:
2725:
2723:
2721:
2720:Text analysis
2717:
2711:
2708:
2706:
2703:
2701:
2698:
2696:
2693:
2689:
2686:
2684:
2681:
2680:
2679:
2676:
2674:
2671:
2669:
2666:
2665:
2663:
2661:General terms
2659:
2655:
2648:
2643:
2641:
2636:
2634:
2629:
2628:
2625:
2613:
2610:
2609:
2607:
2603:
2597:
2594:
2592:
2589:
2587:
2584:
2582:
2579:
2577:
2574:
2572:
2569:
2567:
2564:
2562:
2559:
2557:
2554:
2553:
2551:
2549:
2545:
2539:
2536:
2534:
2531:
2529:
2526:
2524:
2521:
2519:
2516:
2514:
2511:
2510:
2508:
2506:
2505:Free software
2502:
2497:
2490:
2485:
2483:
2478:
2476:
2471:
2470:
2467:
2461:
2458:
2455:
2452:
2451:
2447:
2440:
2435:
2432:
2427:
2423:
2419:
2415:
2411:
2407:
2400:
2397:
2392:
2386:
2378:
2376:9783319245928
2372:
2368:
2367:
2359:
2356:
2351:
2350:
2342:
2339:
2326:
2325:
2317:
2314:
2301:
2294:
2291:
2286:
2280:
2277:
2273:
2271:
2258:
2254:
2248:
2246:
2242:
2237:
2233:
2229:
2225:
2220:
2215:
2211:
2207:
2206:
2197:
2194:
2189:
2183:
2180:
2172:September 20,
2167:
2163:
2157:
2154:
2142:
2138:
2134:
2130:
2123:
2120:
2115:
2111:
2105:
2102:
2089:
2085:
2079:
2077:
2075:
2071:
2059:
2058:www.bisok.com
2055:
2049:
2046:
2041:
2037:
2030:
2027:
2014:
2008:
2005:
1992:
1988:
1982:
1979:
1966:
1960:
1957:
1941:
1934:
1927:
1925:
1923:
1919:
1906:
1900:
1897:
1892:
1888:
1884:
1880:
1876:
1872:
1865:
1862:
1846:
1842:
1838:
1834:
1828:
1824:
1820:
1813:
1812:
1804:
1801:
1785:
1781:
1777:
1773:
1769:
1762:
1755:
1752:
1736:
1732:
1728:
1724:
1720:
1716:
1712:
1705:
1698:
1695:
1679:
1675:
1671:
1667:
1663:
1659:
1655:
1648:
1641:
1638:
1625:
1619:
1617:
1613:
1608:
1604:
1600:
1596:
1592:
1588:
1581:
1578:
1565:
1558:
1555:
1542:
1536:
1533:
1520:
1514:
1511:
1506:
1502:
1496:
1493:
1488:
1482:
1478:
1477:
1469:
1466:
1461:
1457:
1451:
1448:
1443:
1439:
1433:
1430:
1425:
1421:
1414:
1411:
1406:
1402:
1398:
1394:
1390:
1386:
1379:
1376:
1371:
1369:9781683180142
1365:
1361:
1360:
1352:
1349:
1344:
1342:9780943072012
1338:
1333:
1332:
1323:
1321:
1317:
1312:
1308:
1301:
1298:
1293:
1289:
1282:
1279:
1273:
1268:
1265:
1263:
1260:
1258:
1255:
1253:
1250:
1248:
1245:
1243:
1240:
1238:
1235:
1233:
1230:
1228:
1225:
1223:
1220:
1218:
1215:
1213:
1210:
1208:
1205:
1203:
1200:
1198:
1195:
1193:
1190:
1188:
1185:
1183:
1180:
1178:
1175:
1173:
1170:
1168:
1165:
1164:
1159:
1149:
1145:
1141:
1137:
1136:
1135:
1131:
1095:
1094:
1047:
1046:
1042:
1039:
1036:
1033:
1030:
1027:
1024:
1021:
1018:
1015:
1012:
1009:
1006:
1003:
1000:
997:
994:
993:
989:
985:
984:
979:
976:
974:
970:
966:
961:
959:
953:
945:
943:
941:
940:
935:
930:
927:
925:
921:
917:
913:
908:
906:
905:pen computing
902:
897:
895:
890:
886:
884:
879:
874:
869:
867:
863:
855:
850:
843:
838:
831:
826:
816:
804:
799:
790:
789:
783:
781:
779:
774:
770:
766:
761:
760:Crowdsourcing
755:Crowdsourcing
754:
752:
750:
745:
743:
739:
735:
733:
732:dropout color
729:
725:
722:
718:
714:
706:
704:
698:
696:
694:
690:
689:
684:
682:
678:
674:
670:
666:
662:
653:
651:
649:
644:
641:
639:
638:co-occurrence
635:
631:
629:
625:
620:
618:
614:
606:
604:
602:
597:
595:
591:
587:
583:
578:
574:
572:
568:
564:
560:
555:
552:
548:
540:
536:
532:
528:
523:
520:
517:
516:
511:
510:
505:
501:
498:
497:
496:
490:
488:
486:
482:
475:
471:
467:
463:
459:
456:
453:
449:
445:
442:
439:
436:
431:
427:
423:
420:
417:
414:
410:
409:
408:
402:
397:
395:
392:
388:
384:
376:
372:
368:
365:
362:
358:
354:
350:
347:
344:
340:
336:
333:
329:
325:
324:
320:
315:
312:
309:
306:
302:
299:
298:pen computing
295:
293:
289:
287:
283:
282:book scanning
279:
276:
274:
271:
268:
265:
261:
259:
256:
253:
249:
248:Entering data
246:
245:
244:
241:
235:
233:
231:
227:
223:
220:
216:
212:
208:
203:
201:
197:
193:
189:
185:
181:
177:
172:
168:
160:
158:
156:
152:
148:
143:
141:
137:
131:
123:
121:
119:
113:
111:
107:
103:
99:
95:
91:
87:
83:
79:
74:
71:
67:
63:
59:
55:
51:
30:
19:
4138:Hugging Face
4102:David Silver
3776:
3750:Audio–visual
3604:Applications
3583:Augmentation
3428:
3277:Concordancer
3223:
2673:Bag-of-words
2495:
2434:
2409:
2405:
2399:
2365:
2358:
2348:
2341:
2329:. Retrieved
2323:
2316:
2304:. Retrieved
2293:
2279:
2267:
2260:. Retrieved
2256:
2209:
2203:
2196:
2182:
2170:. Retrieved
2165:
2156:
2144:. Retrieved
2132:
2122:
2114:the original
2104:
2092:. Retrieved
2061:. Retrieved
2057:
2048:
2040:ResearchGate
2039:
2029:
2017:. Retrieved
2007:
1995:. Retrieved
1991:the original
1981:
1969:. Retrieved
1967:. Dataid.com
1959:
1947:. Retrieved
1940:the original
1909:. Retrieved
1899:
1874:
1870:
1864:
1852:. Retrieved
1810:
1803:
1791:. Retrieved
1771:
1767:
1754:
1742:. Retrieved
1735:the original
1714:
1710:
1697:
1685:. Retrieved
1678:the original
1657:
1653:
1640:
1628:. Retrieved
1590:
1586:
1580:
1568:. Retrieved
1557:
1545:. Retrieved
1535:
1523:. Retrieved
1513:
1505:the original
1495:
1479:. Springer.
1475:
1468:
1460:the original
1450:
1442:the original
1432:
1423:
1419:
1413:
1388:
1384:
1378:
1358:
1351:
1330:
1311:the original
1300:
1292:the original
1281:
1202:eScriptorium
1148:
1140:
1133:
981:
962:
955:
938:
933:
931:
928:
919:
912:cursive text
909:
898:
891:
887:
873:Latin script
870:
865:
859:
810:
802:
758:
749:Template OCR
748:
746:
736:
727:
726:
710:
702:
692:
686:
685:
657:
645:
642:
633:
632:
621:
617:proper nouns
610:
598:
579:
575:
556:
544:
537:such as the
521:
513:
507:
503:
499:
494:
478:
470:aspect ratio
430:binary image
406:
380:
343:word divider
292:Google Books
242:
239:
236:Applications
224:
215:smartglasses
204:
167:Ray Kurzweil
164:
144:
133:
114:
75:
57:
53:
49:
48:
29:
4286:Categories
4234:Autoencoder
4189:Transformer
4057:Alex Graves
4005:OpenAI Five
3909:IBM Watsonx
3531:Convolution
3509:Overfitting
3234:Topic model
3114:Text corpus
2960:Statistical
2827:Text mining
2668:AI-complete
2566:Asprise OCR
1566:. Ejohn.org
1426:: 46. 1970.
1307:"undefined"
1197:Digital pen
728:Comb fields
699:Workarounds
669:screenshots
559:Google Docs
419:Despeckling
371:printscript
353:printscript
266:in airports
98:text mining
4312:Categories
4275:Technology
4128:EleutherAI
4087:Fei-Fei Li
4082:Yann LeCun
3995:Q-learning
3978:Decisional
3904:IBM Watson
3812:Midjourney
3704:TensorFlow
3551:Activation
3504:Regression
3499:Clustering
2955:Rule-based
2837:Truecasing
2705:Stop words
2586:SmartScore
2331:October 3,
2306:January 5,
2212:(2): 155.
1717:(2): 389.
1660:(1): 146.
1593:(8): 787.
1274:References
1212:Legibility
922:line of a
813:March 2013
624:plain text
586:XML schema
465:connected.
435:scene text
398:Techniques
334:at a time.
211:smartphone
188:LexisNexis
128:See also:
78:data entry
66:mechanical
62:electronic
4158:MIT CSAIL
4123:Anthropic
4092:Andrew Ng
3990:AlphaZero
3834:VideoPoet
3797:AlphaFold
3734:MindSpore
3688:SpiNNaker
3683:Memristor
3590:Diffusion
3566:Rectifier
3546:Batchnorm
3526:Attention
3521:Adversary
3264:reviewing
3062:standards
3060:Types and
2538:Tesseract
2528:OCRFeeder
2513:CuneiForm
2441:Dead link
2385:cite book
2219:1410.6751
2141:0362-4331
1232:Music OCR
1167:AI effect
769:reCAPTCHA
551:Tesseract
547:Cuneiform
426:greyscale
332:character
165:In 1974,
151:microfilm
140:Optophone
60:) is the
4266:Portals
4025:Auto-GPT
3857:Word2vec
3661:Hardware
3578:Datasets
3480:Concepts
3180:Wikidata
3160:FrameNet
3145:BabelNet
3124:Treebank
3094:PropBank
3039:Word2vec
3004:fastText
2885:Stemming
2605:See also
2591:TeleForm
2581:ReadSoft
2576:OmniPage
2498:software
2426:26389649
2262:July 20,
2236:11873638
2146:June 16,
2094:June 16,
2019:June 16,
1997:June 16,
1971:June 16,
1911:June 16,
1845:Archived
1784:Archived
1630:June 16,
1607:42920826
1570:June 16,
1547:June 16,
1525:June 16,
1160:See also
784:Accuracy
742:Graffiti
673:ID cards
665:invoices
391:segments
196:Scansoft
4348:Symbols
4343:Unicode
4148:Meta AI
3985:AlphaGo
3969:PanGu-Σ
3939:ChatGPT
3914:Granite
3862:Seq2seq
3841:Whisper
3762:WaveNet
3757:AlexNet
3729:Flux.jl
3709:PyTorch
3561:Sigmoid
3556:Softmax
3421:General
3351:Related
3317:Chatbot
3175:WordNet
3155:DBpedia
3029:Seq2seq
2773:Parsing
2688:Trigram
2596:VueScan
2533:OCRopus
2063:May 27,
1949:May 23,
1879:Bibcode
1841:8947361
1719:Bibcode
1662:Bibcode
1393:Bibcode
1096:U+245x
1048:U+244x
958:Unicode
946:Unicode
878:cursive
842:n-grams
830:n-grams
803:updated
738:Palm OS
613:lexicon
567:OCRopus
461:script.
413:skewing
375:cursive
357:cursive
305:CAPTCHA
124:History
4163:Huawei
4143:OpenAI
4045:People
4015:MuZero
3877:Gemini
3872:Claude
3807:DALL-E
3719:Theano
3324:(c.f.
2982:models
2970:Neural
2683:Bigram
2678:n-gram
2424:
2373:
2234:
2139:
1854:May 2,
1839:
1829:
1793:May 2,
1744:May 2,
1687:May 2,
1605:
1483:
1366:
1339:
995:
990:(PDF)
934:scanno
920:Amount
894:long s
854:long S
771:. The
679:, and
452:tables
252:checks
178:-type
70:images
4229:Mamba
4000:SARSA
3964:LLaMA
3959:BLOOM
3944:GPT-J
3934:GPT-4
3929:GPT-3
3924:GPT-2
3919:GPT-1
3882:LaMDA
3714:Keras
3373:spaCy
3018:large
3009:GloVe
2523:Ocrad
2232:S2CID
2214:arXiv
1943:(PDF)
1936:(PDF)
1848:(PDF)
1837:S2CID
1815:(PDF)
1787:(PDF)
1764:(PDF)
1738:(PDF)
1707:(PDF)
1681:(PDF)
1650:(PDF)
1603:S2CID
1134:Notes
973:OCR-B
969:OCR-A
924:check
719:, or
717:OCR-B
713:OCR-A
561:OCR,
512:, or
474:scale
341:as a
339:space
328:glyph
321:Types
192:Xerox
4153:Mila
3954:PaLM
3887:Bard
3867:BERT
3850:Text
3829:Sora
3138:Data
2989:BERT
2518:GOCR
2422:PMID
2391:link
2371:ISBN
2333:2008
2308:2014
2264:2023
2174:2018
2148:2023
2137:ISSN
2096:2013
2065:2021
2021:2013
1999:2013
1973:2013
1951:2013
1913:2013
1856:2015
1827:ISBN
1795:2015
1746:2015
1689:2015
1632:2013
1572:2013
1549:2013
1527:2013
1481:ISBN
1364:ISBN
1337:ISBN
965:MICR
939:typo
767:and
721:MICR
646:The
594:hOCR
582:ALTO
549:and
472:and
450:and
284:for
171:font
108:and
3894:NMT
3777:OCR
3772:HWR
3724:JAX
3678:VPU
3673:TPU
3668:IPU
3492:SGD
3170:UBY
2414:doi
2224:doi
1887:doi
1819:doi
1776:doi
1727:doi
1670:doi
1595:doi
1401:doi
971:or
942:).
628:PDF
411:De-
373:or
355:or
330:or
219:API
176:CCD
155:IBM
64:or
58:OCR
52:or
4314::
2420:.
2408:.
2387:}}
2383:{{
2266:.
2255:.
2244:^
2230:.
2222:.
2210:19
2208:.
2164:.
2135:.
2131:.
2073:^
2056:.
2038:.
1921:^
1885:.
1875:29
1873:.
1843:.
1835:.
1825:.
1782:.
1772:17
1770:.
1766:.
1725:.
1715:40
1713:.
1709:.
1668:.
1658:13
1656:.
1652:.
1615:^
1601:.
1591:12
1589:.
1424:12
1422:.
1399:.
1389:90
1387:.
1319:^
1146:2.
1138:1.
1081:⑊
1078:⑉
1075:⑈
1072:⑇
1069:⑆
1066:⑅
1063:⑄
1060:⑃
1057:⑂
1054:⑁
1051:⑀
1043:F
975:.
967:,
780:.
751:.
715:,
683:.
675:,
671:,
667:,
663:,
603:.
506:,
202:.
157:.
112:.
104:,
88:,
3407:e
3400:t
3393:v
3328:)
3051:,
3020:)
3016:(
2646:e
2639:t
2632:v
2488:e
2481:t
2474:v
2428:.
2416::
2410:4
2393:)
2379:.
2335:.
2310:.
2238:.
2226::
2216::
2176:.
2150:.
2098:.
2067:.
2042:.
2023:.
2001:.
1975:.
1953:.
1915:.
1893:.
1889::
1881::
1858:.
1821::
1797:.
1778::
1748:.
1729::
1721::
1691:.
1672::
1664::
1634:.
1609:.
1597::
1574:.
1551:.
1529:.
1489:.
1407:.
1403::
1395::
1372:.
1345:.
1040:E
1037:D
1034:C
1031:B
1028:A
1025:9
1022:8
1019:7
1016:6
1013:5
1010:4
1007:3
1004:2
1001:1
998:0
815:)
811:(
805:.
454:.
363:.
300:)
56:(
20:)
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.