Speech is now not confined solely to the human race, or even living beings on a broader perspective. It has extended to the World of Technology; specifically Computers and Software. Speech Recognition Technology [SRT], little known until recent times, is an area that has been constantly growing over the years. SRT has opened new windows of panoramic dimensions to human kind.
Speech Recognition Technology – Defined:
A technology wherein a machine or program identifies spoken words and phrases and converts them to readable format. In more simpler terms, voice/speech is converted to text format.
Speech Recognition Technology, as we know it today, did not just bloom overnight. It is the work of over 30 years. The very first implementation of Speech Recognition was designed and displayed by IBM during the 1964 New York World’s Fair. It was called the IBM Shoebox; and typically sized too. This device only recognized spoken digits from 0 to 9. Over the years, Software companies have delved upon the benefits and the profits that can be reaped from developing SRT, and they have come a long way since. Now many noted companies have come up with different versions of this technology, offering various features.
While Speech Recognition has progressed vastly, the downside is that it has been a long and laborious road this far. It calls for training and updating of Software to work flawlessly; and it is currently far from perfect. There is so much more to go before it can be said that SRT is without any imperfections.
Speech Recognition Technology – In Health Care and Medical Transcription:
An area where this technology has caught on to a great extent is the Medical Transcription Industry. Speech Recognition is of two types: Front End and Back End.
- Front End: As the doctor dictates into the machine, the words are automatically generated and displayed. This text can be edited directly, the report finalized, and signed by the doctor immediately. But there are few Front end users due to time constraints faced by doctors and practitioners.
- Back End: Also called Delayed or Deferred Speech Recognition. The written draft is generated by the Software and sent to the Medical Transcriptionists for editing and proof reading. This is more convenient for doctors as it requires less time spent reading the proofs.
- EMR: Speech Recognition can be applied in Electronic Medical Record systems of hospitals, clinics, etc. Searches, queries, and even filling up of forms can be made simpler and faster with voice rather than using the keyboard.
Speech Recognition Technology can be advantageous to both Medical Transcriptionists as well as doctors in a number of ways. A few advantages are:
- There is not much typing involved as the first draft is generated by the software
- Requires only editing of the machine-generated text
- Faster turnaround of reports
- Saves time in completion of reports
- Stat reports can be completed and returned quicker than by using the conventional method of typing
Being a technology that has not entirely been perfected, SRT has a few drawbacks.
- It cannot fully distinguish words spoken with heavy accents
- Homophones (words which sound the same but may have different spellings and meanings) can cause spelling errors
- Incorrect grammar is bound to occur in the generated texts
- Punctuation rules are not likely to be followed at all times
- Multiple speakers cannot be clearly differentiated
- Background noise and disturbances may cause misspellings
Question of the Hour:
A niggling thought in the minds of many Medical Transcriptionists is whether Speech Recognition Technology poses a threat to their livelihood. This is a baseless fear, keeping in mind the drawbacks of SRT. No report will be complete until a human eye has scanned it. Machine-generated documents can only be 60%-70% error-free. It demands well-trained and professional Medical Transcriptionists to read through and edit the reports. A mistake, even a tiny one, can result in unpleasant situations for many. And one cannot rely on mere machines to do the job 100% without errors. In conclusion, Speech Recognition Technology is certainly one of a kind. While it has not yet reached the peak of perfection, it is still quite an aid to the healthcare industry, specifically Medical Transcription. And there is certainly no danger of software or machines or even technology entirely taking over man’s work; technology is just merely helping out!