LibGuides
Examples
This example from IASL 2023-24 and 2024-25 Graduate Assistant Alice Tierney-Fife shows how the camera function on the Google Translate app can be used to overlay a translation onto whatever you hold your camera up to -- including websites, signage, books, museum labels, archival documents, and more! This is a great first-step resource when you need to get the gist of something quickly, or kickstart your own translation work.
Transkribus is a great tool to use for text recognition.
Miwo is an excellent option for Japanese resources, including calligraphy and handwritten texts.
General Tools for Many Languages
Many of the tools cited below can be good options across many language families. Google Translate is an easily accessible option; others include ABBYY FineReader, ReadMe, and Tesseract, which are detailed on the OCR page as well.
- ABBYY FineReaderThis OCR program converts image documents such as photos, scans, PDFs, and screen captures into file formats that can be edited. You can convert your photos into Word, Powerpoint, or Excel documents, into searchable and editable PDFs, into HTML and plain text files, and more! And of special interest to area studies scholars, Version 15 supports text recognition in 192 languages, with built-in spellcheck for 48.
- Google TranslateChoose your language, select the photo icon, and take a photo of the text you wish to translate! The Google Translate app will scan the text and translate it over the image for you. It works with many languages and scripts, and is an incredible starting point for your translation and transliteration work. If you are an Android phone user, you can also easily access these features through Google Lens.
- ReadMeWith a free option, ReadMe can be a great choice for working with APIs and OCR technology. It is customizable and works well with languages.
- TesseractTesseract is a useful multilingual OCR tool that converts image documents into PDF or Word documents, which you can then edit or search.
Unlike some other tools that are available, Tesseract requires some basic coding knowledge.
This link connects to the GitHub Tesseract code page.
African and Middle Eastern Languages
- Article: Abdullahi Tsanni, "This company is building AI for African languages"This article from the MIT Technology Review (2023) gives an interesting overview into the work and mission of Lelapa AI and Lesan.
- Article: J. Hocking and M. Puttkammer, "Optical character recognition for South African languages," 2016 Pattern Recognition Association of South Africa and Robotics and Mechatronics International Conference (PRASA-RobMech), Stellenbosch, South Africa, 20Abstract:
Optical Character Recognition (OCR) is an essential technology in the digitisation of printed media. Many OCR engines are language-specific and available for common languages such as English and other European languages, but less so for smaller languages such as Tshivenda and the other South African languages. In this paper, we describe the process of training OCR engines for South African languages using Tesseract, and compare the accuracy of these engines with the accuracy of using an English engine on the other South African languages. We find that our language-specific engines achieve a high accuracy, with 50% less errors as when using an English engine. - Article: Will McBain, "AI tools open up a world of African languages"This article from African Business (March 2023) gives a good overview of AI, language learning, and OCR technology that is currently enabled for African languages, as well as African business and technology ventures that will continue to expand and create resources and opportunities in the future.
- Convertio OCRWhile only free for the first 10 pages, Convertio is still a decent choice that works with a range of languages, including Afrikaans, Arabic, and Swahili. You can select multiple languages in use in one document, and upload in multiple formats. You can also choose the output format from a number of options and save your converted files to Dropbox or Google Drive.
- Ghana Natural Language ProcessingGhana NLP is an open-source movement that provides smart keyboards for writing in African languages. Their current projects are committed to developing natural language processing technology dedicated to Ghanian languages such as Twi and Ewe, which are often left behind by global companies such as Google Translate.
- Lelapa AILelapa AI is a company working to create AI language models that work with African languages. Their tool Vulavula converts voice to text and works with English, Afrikaans, isiZulu and Sesotho -- with "other official South African languages and a bunch of Sub-Saharan languages coming soon"!
- Lesan AIThis AI startup is working to develop translation tools for Ethiopian languages. Their translation service currently works with English, Tigrinya, and Amharic.
- NanonetsNanonets is a great tool for over 200 languages, which includes Arabic. It uses AI-based OCR software to recognize text, including with useful items such as passports, receipts, and bills. It is also very useful for processing documents. It has a reported 95% OCR precision rate. It is also a good option for Russian, as detailed below.
- Sakhr OCRSakhr is an OCR tool that is specific to Arabic and Arabic-script text recognition, with a high precision rating for Arabic, Farsi, Pashto, Jawi, and Urdu. It recognizes scanned documents and images.
Western European Languages
- Amazon TextractAmazon Textract can be another good option for working with European languages. It extracts printed text and handwriting, along with tables and other structured data, from documents and images.
- Convertio OCRDetailed in the "African and Middle Eastern Languages" section above, Convertio also offers options for many Western European languages.
- TranskribusTranskribus is a super useful AI-powered platform designed for text recognition, transcription, and searching for historical documents. It supports a wide range of European languages, and options for Arabic and Japanese are in development.
Asian Languages
- Dharmamitra TranslatorThe Dharmamitra Translator is a tool that uses AI to translate Buddhist texts from Sanskrit, Tibetan, and Classical Chinese into various modern languages.
- Miwo AppReading くずし字 (cursive script) is one of the most challenging aspects of examining pre-modern Japanese texts. However, AI tools have been increasingly helpful in this area. Mio is an application that uses AI to capture cursive writings through a phone camera and converts them into modern characters.
- 古籍酷 GJ.coolThis platform supports the OCR, auto punctuation, and translation of pre-modern Chinese books.
- 古联 OCR/Gulian OCRThis tool is specifically designed for the OCR of pre-modern Chinese woodblock prints. It can also identify comments and annotations to prevent them from being mixed with the main text during the conversion from images.
Slavic Languages
- Convertio OCRDetailed in the "African and Middle Eastern Languages" section above, Convertio also offers OCR support for a number of Slavic languages.
- i2OCRi2OCR is a free Russian OCR platform which works well for formatting, changing, translating, and searching texts from images, in Russian and other languages. However, it has limitations in what documents it can use (only images) and comparatively low-quality precision, formatting, and result time.
- NanonetsNanonets is a great tool for over 200 languages, which includes Russian and other Cyrillic-script languages. It uses AI-based OCR software to recognize text, including with useful items such as passports, receipts, and bills. It is also very useful for processing documents. It has a reported 95% OCR precision rate, so is a more reliable choice for Russian than many other options.
- OCR ConvertThis free online tool offers conversion technology for a number of languages, including many Slavic languages. You can upload a document file or URL link, and convert it to a txt file.
- Text Scanner RussianThis Android app can be used to translate text from your camera images. While the precision is not yet as high as some other options, the app is being improved and can be a good choice when you are on the go, or need a quick general idea of what you are looking at while doing research.
Online Workspaces for Asian Language Primary Sources
Online databases are excellent sources of primary textual and pictorial materials. Many researchers utilize these databases, searching and copying text for their studies. To avoid repetitive searching, it's crucial to effectively save and organize primary sources relevant to our research topics. Several information organizations have developed useful online platforms to aid researchers in creating personal databases and collaborating on projects concerning pre-modern texts. Highlighted below are platforms particularly suited for researchers in East Asian studies.
- DocuSkyDocuSky is an online platform designed for personalized organization and analysis of texts, particularly those in Sinitic scripts. It's conveniently connected to numerous major databases and comes equipped with a variety of digital humanities tools, including punctuation, markup, aligned reading, textual analysis, and data visualization.
- Japan SearchOperated by the National Diet Library, Japan Search is a national platform that aggregates metadata of digital resources across various fields and information organizations. Currently, it is linked to 230 databases. The "My Note" feature allows users to save and annotate their favorite items, galleries, and search results.
- MARKUSMARKUS, a reading and text analysis platform, is extensively used by scholars in Chinese studies worldwide. Its keyword tagging feature aids in named entity recognition and data visualization, making it highly applicable in the field of digital humanities.
- Research Environment for Ancient Documents (READ) FeaturesREAD is an open-source platform designed for epigraphic and manuscript research. It offers a range of features that aid in transcribing, researching, understanding, analyzing, and publishing scholarly editions and studies. Its specialized tools make it particularly useful for working with the Gāndhārī or Sanskrit scripts.
- The Ten Thousand Rooms ProjectThe Ten Thousand Rooms Project is a workspace that allows researchers to collaborate on transcribing, translating, and commenting on texts in East Asian languages. These texts can be uploaded and modified as images.
Research
Yue Lu, F. Y. & Chew Lim Tan. (2004). Chinese Word Searching in Imaged Documents. International Journal of Pattern Recognition & Artificial Intelligence, 18(2), 229–246.
Khan, Muzammil, Arif Ur Rahman, Arshad Ahmad, and Sarwar Shah Khan. (2022). A Content-Based Technique for Linking Dual Language News Articles in an Archive. Journal of Information Science, 48 (1): 57–70.