Multilingual Translation Technology Centre

An Introduction to Machine Translation at the MTTC

Machine translation (MT) is the application of computers to the task of translating texts from one language to another. The RCAT's Multilingual Translation Technology Centre (MTTC) offers a complete, no-cost service to scan and translate your documents to English or from English to one of the target languages.

There are essentially two steps to translating a printed document: scanning and translation, both of which are done at the same workstation. Printed documents can be scanned as loose pages, converted to electronic text that is recognizable by the translation software, and then translated into the desired language using our software. For detailed instructions on translating your documents using our equipment please visit the Instructions page.

Languages Supported

At the moment, our OmniPage 15.0 OCR scanning software recognizes Latin and Cyrillic characters and supports the following languages: English, French, Spanish, Italian, Portuguese, German, Dutch, Swedish and Russian. The SYSTRAN 5.0 translation software supports all the languages above, as well as Japanese, Korean and traditional and simplified Chinese. However, unless your document is in one of the languages supported by the OCR software, only those documents that are already in electronic format may be translated.

Optical Character Recognition

Optical character recognition (OCR) is the electronic translation of images of printed text (such as a page of a book or journal), captured by a scanner, into machine-editable text. If the document you have is already in an electronic text format (such as .doc or .rtf) it can be translated directly. If it is a printed document you will first need to scan it and convert it to text using our OCR software.

Accuracy rates for the recognition of Latin-script, typewritten text now exceed 99%. Because the accuracy is directly influenced by the quality of the document, it is important to manually review the scan for formatting errors, which often occur with complex document layouts, before attempting to translate it. For more information of scanning your documents please visit the Instructions page.

Translation Accuracy

While character recognition can now achieve over 99% accuracy, and machine translation has been the subject of research for over 50 years, this process is not faultless and requires human intervention at various stages. Because of the complexity of structure and meaning in human languages, machine translation is a great tool for gaining an overall understanding of a foreign-language text, but it cannot substitute professional translation by a human being.

Complex document structure that involves tables, small and various fonts and symbols may impact the quality of the scan, so that the final result often depends heavily on the quality and legibility of the original document.

Extra Support

The MTTC is at your disposal as a self-serve system. If this is your first time using it our you would like to receive additional help, please contact us to set-up a time for one-on-one training at info.rcat@utoronto.ca.

Hours and Location

The Multilingual Translation Technology Centre is located at the:

Resource Centre for Academic Technology
Robarts Library, 4th Floor
130 St. George Street
University of Toronto
Toronto, Ontario M5S 1A5

Our hours of operation are 9 am - 5 pm on Monday - Friday. Please visit the Computer Booking page for available times and to reserve the MTTC. Our service is free of charge, but it is available to faculty, staff and students of the University of Toronto only.