Scientists have created a software to read and transcribe old documents

There is no praise high enough for the famous English engineer Isambard Kingdom Brunel. He designed revolutionary transatlantic steamships, dockyards, bridges and tunnels. Brunel is still considered to be one of the most important and innovative engineers in history and he definitely was the central figure of the Industrial Revolution. But boy his handwriting was bad. But fear not – artificial intelligence is here to help.

Isambard Kingdom Brunel was one of the most important engineers in history, but his handwriting is challenging.Image credit: Robert Howlett via Wikimedia

Isambard Kingdom Brunel died relatively young – he was just 53 when he suffered a stroke and died. He was a heavy smoker and always hanged out in industrial environment filled with air pollution. His legacy was not only the ships and infrastructure that he’s created, but also a huge pile of diaries and documents, containing a lot of valuable information and, probably, creations that never materialized.

Brunel’s handwriting is challenging to say the least. It is not that it is impossible to read, but it does take a lot of time to decipher and transcribe it. In fact, it is so challenging that scientists decided to create an AI system just to read it – Transkribus learns the specific quirks of an author’s writing, compares it with examples which have already been transcribed, and presents the text in an easily accessible format.

The Transkribus was developed by scientists from all over Europe and it is truly genious. However, even it can currently read just 65% of Brunel’s words. Just to start it needs a minimum of 15,000 manually transcribed words to be presented to it, so that it could start making sense of what the texts are saying. It is already very accurate in the places where it can actually read.

Of course, Transkribus software will be very helpful in many areas, not just transcribing old Brunel’s documents. It will help digitizing a lot of vintage scientific material, which will inevitably lead to new discoveries.  Professor Philip Schofield, co-author of the study, said: “There are millions upon millions of pages of handwritten documents in archives around the world. Transkribus has the potential to revolutionise the way archivists and researchers read, transcribe, process and mine historical documents”.

Old texts are full of secrets that are yet to be solved. And they may contain very valuable information. However, because they are so unique, these documents are mostly inaccessible to majority of scientists. Transcribing and digitizing them would benefit science community and the world as well.

 

Source: UCL


<!–

Comment this news or article

–>