A blind person empowered to catch the correct bus. A non-English speaking Chinese tourist navigating their way around Australian place names with ease.
Getting rid of the office scanner in favour of a simple function on the Smartphone.
All of these possibilities of the near future are the focus of current Griffith research to be featured at the 10th IAPR International Workshop on Document Analysis Systems.
Held March 27-29 at the Crowne Plaza, Surfers Paradise, the bi-annual research workshop is hosted by Griffith University and will feature a range of research including, but not limited to, document image analysis and handwriting recognition technology.
Associate Professor Michael Blumenstein from Griffith’s Science, Environment, Engineering and Technology (SEET) Executive said the university holds the critical mass of this type of research within Australia and that it is natural for it to host the event.
“This year we have a magnificent line up of academic researchers and industry experts from various universities and companies around the world, including a special showcase of our own leading research here on the Gold Coast.”
The creation of a prototype for a video-based document processing system which can be used in a range of applications is the focus of a research paper to be presented by Griffith PhD student Nabin Sharma.
“My research will give an insight into the future of how this technology can assist
various user groups. For example, it is envisaged that a blind person could easily
use video technology on their smartphone to capture relevant bus information.
“That would then be translated into a signal such as a beep or vibration which would tell the user of the correct bus’s whereabouts.”
The implementation of Optical Character Recognition (OCR) technology within Smartphones could also prove a convenient function for Smartphone users looking to efficiently scan documents.
“This could potentially be a commonplace technology within Smartphones over the next four years, and in the process cause the traditional scanner with all its bulkiness and lack of portability, to become obsolete,” Mr Sharma said.
“This will have real benefit to users, not only in terms of the convenience, but also
in the way that it will be able to take the textual contents of certain, perhaps
delicate and difficult-to-read historical documentation, and recreate it into a more
editable and searchable format.”
“The whole point of document analysis system technology is to be able to extract information without the need for annotation or any prior human interaction,” said Associate Professor Michael Blumenstein.
“If systems are able to recognise, verify and store for example, hand-written forms in seconds, which traditionally may have taken a significantly longer time, then the
productivity and efficiency gains for government, business and research could be
absolutely enormous.”
Dr Samy Bengio, a research scientist in machine learning from Google, will be leading the keynote address at the event. It is tentatively titled “Learning A Semantic Space: From Image Annotation to Music Similarity”.