Paragon PenReader

PenReader is the application for intellectual recognition of hand-written input. A system works for 38 languages.

PenReader will recognize different options of handwriting in different languages, in independence of the place and a method of introduction on the screen of the mobile device. PenReader is compatible to all applications at the level of the operating system. At choice four different modes of recognition are offered to the user: conjoint, separate, letter by letter and intellectual, the last is set by default. The program will recognize letters, digits, arithmetic and punctuation characters and also allows to work with macroes. PenReader allows to automatically fill up the user dictionary base.

The SDK window for developers (2014)

PenReader has the expanded system of settings thanks to what it is possible to teach to recognize the application all features of individual handwriting. Settings, in particular, allow to set a recognition delay time (a time frame which should pass from the moment of a lead over the screen of the mobile device without the subsequent contacts), to define active languages of input, to configure width and color of a feather and many other things. For acceleration of process of recognition and reduction of number of spelling errors the program offers options for end.

Functionality

Four input modes:

• "Letter by letter",
• "Conjoint",
• "Separate",
• "Intellectual".

• convenient text entering on devices with small screens: the Letter on a Letter mode is developed especially for small screens. It is possible to enter the whole word, drawing letters the friend on the friend - characters will be recognized as one word.

• the "Collect a Word in Parts" function allows to enter a part of a word on the small screen, and PenReader will offer options of words at choice.

• fast and exact words recognition by means of MegaGraf technology

• the user dictionary – a possibility of automatic adding of words in the user dictionary. Hints in the lower part of the screen offer options of writing of words.

• function of spelling checks the text for errors at once (writing is verified with the spelling dictionary).

• the analysis of the next characters increases input accuracy in Conjoint and Separate the modes.

• special characters ($, @, &, etc.) are displayed on the screen keyboard now. Long clicking the button opens access to other characters, in addition to the basic.

• for convenient use of setup are grouped in thematic sections, descriptions under headings of settings and some points are added.

• the application occupies the small amount of memory of the device.

PenReader for iOS

On January 14, 2016 the Paragon Software company announced availability of PenReader to users of devices running iOS.

The PenReader technology was used on the mobile devices Newton Apple 20 years ago. For January, 2016 PenReader has the minimalist interface and implements sensing technologies, processings of hand-written characters.

The Paragon Software company announced release of the complete version of the software application of recognition of hand-written input of PenReader for iPhone, iPod touch and iPad. Earlier software was offered for devices on Android, Windows Mobile, tablet PCs and desktop computers with Windows OS.

Since January 14, 2016 PenReader is compatible to all applications at the level of operating system iOS 9, including the browser Safari, SMS, messengers, social networks and e-mail clients. PenReader will allow to simplify work with media: Medium, WordPress, Periscope and others.

We are happy with result – the application which in essence the system utility, should be pleasant to all users of IOS, especially to those who use Pencil Apple and iPad Pro. We tried to make convenient even more convenient, at the same time without limiting the user in opportunities of setup and personalisation. Special attention was paid to dictionary technology – each language requires separate completion, and thanks to our experience in the development area of dictionary applications of Slovoed, we achieved impressive result – for example, for quality improvement of the Russian-language version we analyzed 283,000 texts and more than 400 types of handwriting of native speakers. Nikolay Ablyazov, head of department of development of SHDD Paragon Software

Several dictionary packets are available to users:

• Cyrillic pack is Russian, Belarusian, Bulgarian, Kazakh, Serbian and Ukrainian

• Latin pack is English, Albanian, Catalan, Croatian, Czech, Danish, Dutch, Esperanto, Estonian, Finnish, French, German, Bulgarian, Icelandic, Italian, Latvian, Lithuanian, Maltese, Norwegian, Polish, Portuguese, Romanian, Slovak, Slovenian, Spanish, Swedish, Turkish,

• Full pack is all supported languages

Arab, Hebrew, Greek and Georgian are available.

History

• In 1997 the Paragon Software company developed own systems of recognition of hand-written input for the first PDA and mobile devices.

In 1997 PenReader became available to users of mobile devices of Apple company - Newton Apple.

We for the first time tested possibilities of use of technology for mobile devices on Newton Apple, and now, nearly 20 years later, PenReader is available to iPhone, iPod touch and iPad – it is rather symbolical. Text recognition belongs to subtasks of creation of artificial intelligence, I devoted to it the last 50 years. Right at the beginning it seemed that it is quite simple task that in two-three months it is possible to teach to perceive the machine the handwritten text, to actually understand the person. But on it decades left, and these are thousands of hours of laborious work of hundreds of scientists worldwide and millions of investments. It turned out not at all, we are glad that we are among those who managed it. In 1997 when the first prototype of the program was ready, we also could not imagine that it only the beginning of a long way. Imagine that the input flow of hand-written input has no formalized restrictions. It is also the main complexity in improvement of sensing technologies and processings of hand-written input. But we found a way out – we relied on theoretical learnability of the program, on possibilities of our dictionary technology which we constantly develop and we improve.