Think Before You Speak!

The future of speech-recognition software is currently on hold. Despite the prevalence of speech-recognition software, few breakthroughs in this field have been made in the last ten years. More processing power for computers during this time has not translated into better software applications.

The biggest issue behind the delay concerns pronunciation. Words can be pronounced dozens of different ways. Requiring that a computer understand every pronunciation for most words would entail a large database. Identifying every possible pronunciation of a word, finding people who use that pronunciation, recording the usage, and storing said pronunciation on an accessible database would require a massive investment of time and money. 

Once this challenge to speech-recognition software and others are overcome, the medical community stands to benefit significantly.  Speech-recognition software would make compiling fully computerized medical records a reality by allowing for direct capture of physicians’ notes. If speech-recognition software with a high accuracy rate was implemented, doctors could see more patients annually, decrease costs, and increase accuracy of record keeping.

while I agree that this

while I agree that this technology is greatly needes, I don't think it will improve the medical profession in the way you suggest. My wife has work for several years in the medical profession and Many doctors are already double booking an seeing more patients in a day than they can handle. the quick in and out system that many physicians are having to use to break even on insurance costs, that leads to mis-diagnosis. over all a great post. keep it up

One intermediary step there

One intermediary step there would be to moving to electronic patient records in the first place. There's a lot of patient information stored in all sorts of written forms. That said, speech and handwriting recognition can make that process faster. Happily, funding for switching over to electronic records is in the stimulus.

Oh, and the link doesn't work. I think it's confused by a '('