Speech to Text, Automatic Punctuation

[kjv1611]

I've been pretty interested (though not made very much use of, to date) the whole speech recognition software area for a good while.

So along that line, I've tinkered a little with Dragon Naturally Speaking, but never really made good use of it.

Along those lines, I got to thinking, wondering, about the punctuation part of things. If current speech recognition cannot automatically add in the periods, commas, paragraphs without a command from the speaker, then it's not all that great for pre-recorded speech.

So, I wonder firstly, is there a commercial product out there that effectively adds in proper punctuation. So far, I didn't find any. However, I did come accross a couple of interesting finds:

1. A US patent listing from last month along the same line:

http://www.patentstorm.us/patents/7580838/description.html

2. A UK patent listing for about the same thing, but no detail nor date given:

http://www.freepatentsonline.com/EP1422692.html

3. A paper on the same thought/issue, or at least part of it:

http://www.assta.org/sst/SST-94-Vol-ll/cache/SST-94-VOL2-Chapter12-p2.pdf

So, along those lines, I've got to thinking. If a person wanted to make some sort of attempt at developing their own system to automatically punctuate unpunctuated text, then:
1. What would be the best programming language, or perhaps what would possibly be the top few choices?

2. Would a relational database structure for storing dictionaries be best, or would that inevitably be too resource intensive, and therefore too slow?

Well, I thought that since I'm talking about language, and thinking more in a general sense about technology, it'd make for a fine discussion here - maybe.

Does anyone have any thoughts, references, opinions, on this concept?

--

"If to err is human, then I must be some kind of human!" -Me

 

[CajunCenturion]

==> Does anyone have any thoughts, references, opinions, on this concept?
I have a few. I did my Masters research on natural language interfaces to databases and automated semantic processing of queries. Although my research dealt with written text, the very first step is still the same. You must first, as a developer in this arena, have a very thorough working knowledge of all parts of speech, sentence structure, and the interaction between the two. You must fully understand syntax, punctuation, and grammar rules before you can begin to address the big catch-22 of this problem.

--------------
Good Luck
To get the most from your Tek-Tips experience, please read
FAQ181-2886
As a circle of light increases so does the circumference of darkness around it. - Albert Einstein

 

[kjv1611]

Any other thoughts, references, opinions?

--

"If to err is human, then I must be some kind of human!" -Me