Is there an Ubimedia effect on document capture?

Is there an Ubimedia effect on document management?

Until the recent IT history, the frontier between structured and non-structured information was inconceivable. Nowadays, there is a grey area which is expending. In this area meta-information can be added to non-structured information automatically. Some tools are even able to transform non-structured informations into structured informations.

Knowledge maangement

This trend has an effect on document management system, internet and opens a range of new opportunities on the future Ubimedia world. But is it the beginning of new era ? Were is the state of the Art ? What is missing ?

In this article, you will find some experiences, hints and understanding how can hand writing recognition, image recognition, speech recognition improve your daily IT usage now and in the future.

State of the art in Document Recognition

In the past if you want to use hand writing recognition or image recognition, it was delegated in some specialized applications. Nowadays, Windows 7, Apple, Android phone and tablets are providing tools for writing, image and sound recognition included out of the box.

I have tried the beginning of sound recognition in the 90’s with a soundblaster program and now windows 7. Once you have done the learning in Windows 7, the recognition rate and quality improvement are impressive. The CPU power available permits to handle efficiently word processing in parallel with speech recognition on PCs.

For the hand written recognition, the recognition rate and quality have also improved a lot compared to the 90’s. I will not compare hand written and speech recognition rate because I think it is a non-sense. Those interfaces should be better joined seamlessly. What else ?

An efficient documentation tool for Students Microsoft OneNote

I’ve done most of my ESCEM MBA using Microsoft OneNote. The experience on several months daily experience is good. You can classify your notes on many dimensions. You can integrate speeches, videos, presentation in it. You can also create comments on powerpoint document. You can do text recognition on scans to do full text search on every training materials. I like also the integration with internet explorer to export web pages directly in OneNote. But the internet usage of this tool is limited due to the size of the files in OneNote and Microsoft cloud file size limitation to 100 MB which makes it unusable.

Microsoft OneNote strucutures your documentation

Microsoft OneNote


The project sharing tool from Google documents

During ESCEM MBA, we add several experience of parallel work all together on one documents. It is rather impressive and saves a lot of time. The ability to combine simultaneous writing is boosting productivity and creativity. Authors have of course to desire to share their work and benefits. It requires an open mind and humility to be only one actor of the overall process.

We have seen two approaches of knowledge creation and content management. The experience let us a taste of a single cup of coffee, so what else?

Other way to automatise document entry?

Some specialist like Philips have adapted their software from a standalone usage to a process oriented usage from attorney to  secretary for example Field Fisher Waterhouse. Many companies can benefits with mixed technology approach to non-structured information. The idea is not only to provide a content management tool but to boost the full document production chain from creation, collaboration to retrieval integrating best of breed collaboration and content creation technology.

Speech recognition drawbacks?

On a daily basis, speech recognition looses its efficiency in noisy environments. It is due to the implementation of adaptive noise cancellation algorithms. Best practice are currently stucked in bi-entry sources approach : one for the speaker and one for the room/car noise cancellation. In the future, it will joined with subject recognition and 3D noise cancellation to take into account the room audio characteristics.

Hand writing using e-Pens and tablets are starting to be very efficient in text recognition. But in mixed usage with scheme, graphics and texts, the results are poor. On tablets, the CPU power is currently limiting user experiences but this point will soon be a past statement with dual core CPU generalisation.