Recently, Google updated its engine on many smartphones in order to add the new technology of Google Lens to its search engine, thereby allowing the phone to recognize images and their contents and use this information within Google searches. This post will explore the functionally of this new feature with Syriac texts. Early results are convincing especially in terms of the automatic recognition of printed Syriac texts. Here I would like to share my preliminary tests with you and how this new technology can be employed for those who use Syriac in their research.
The location of Google Lens in the updated google search engine.
My first test was based on curiosity. I wanted to see what Google Lens would do with book covers such as “The Syriac Orthodox in North America.” Google Lens recognized that the author was G. A. Kiraz and directed me to his profile and other printed books.
The cover of The Syriac Orthodox in North America by G. A. Kiraz.
Information that Google provided based on the book cover.
The second test I performed was to see what Google Lens would do with printed text from a Syriac version of the Bible. Surprisingly, Google Lens could recognize the text and understand that it was written in Syriac. Furthermore, it offered a transcription, which could be shared with a linked computer. Although the automated transcription required some editing in order to remove some symbols which had creeped into the text during the process of transcription, it was generally quite readable. I thought this was simply fantastic! I have long been concerned with how to perform OCR on Syriac texts (see my previous piece on the topic). Once a time-consuming task, we can now take images with our smartphones and read a transcription of the text in the scripts and fonts that we like within a few seconds!
Google Lens at work on a Syriac verison of the Bible.
Testing Google Lens on the Peshitta.
The final test was to use Google Lens to transcribe the first chapter of the Syriac Gospel of John. After taking a photo it is possible to transcribe the text digitally and then copy it to a laptop that is using the same Google account. Thereafter, one can paste the transcribed text into Microsoft Word or other programmes. Performing these tasks on the Gospel of John worked well. I wondered if it is possibly time to say goodbye to manual transcription, particularly if we are concerned with searching for some specific Syriac words inside lengthy texts (as discussed in my previous piece).
Copying a transcription to one’s computer.
The transcription when pasted into Microsoft Word.
Despite all this, there are still some shortcomings with Google Lens. For example, the linking between transcribed Syriac texts and other online corpuses and databases is weak. Nevertheless, carefully selecting our textual images and running a few tests can enable us to reach some excellent results. I tested Google Lens with the famous Syriac verse commonly attributed to St Ephrem the Syrian “ܐܠܗܐ ܗܒ ܝܘܠܦܢܐ ܠܐܝܢܐ ܕܪܚܡ ܝܘܠܦܢܐ,” and this produced some good results.
Linking to other databases and corpuses.
The tests that I have performed so far illustrate how digital intelligence can function very well while dealing with Syriac texts photographed and processed with Google Lens. The question now is: Do we still need to look for OCR software and applications to transcribe and recognize Syriac texts from images? Perhaps. When used with Syriac texts Google recognition technology is not perfect in terms of accuracy. However, Google offers the possibility of linking various mega databases and corpuses to a transcribed text, and this is absolutely marvelous!
An example of using Google Lens with MSS. The results are primitive, but functional.