Scribe Speech Recognition Overview

 

Introduction

Scribe offers two types of voice recognition (VR) processing for medical transcription – Interactive and Batch.  Each of these methods is supported by the same Scribe proprietary tools for VR processing customization.

 

Voice Recognition Overview

VR is an evolving technology.  While it gets significant press, the reality is that it is available today, and useful depending on individual preferences.  VR technology is based on three components:

 

VR Engine – this is the software that actually processes and translates speech into text.  These are similar.  Scribe currently uses the Microsoft Engine.

 

VR Vocabulary – this is a dictionary of sorts that allows the VR Engine to figure out how to translate your speech to specific words.  Like any dictionary, it is only as good as its contents.  If a word is not in the dictionary, the VR Engine will not recognize it properly.  Scribe has developed its own VR dictionary which includes all of the commonly used terms in Stedman’s medical dictionary.  In addition, different that other providers, you can easily add terms to our vocabulary through our simple web based interface.  This allows for continued customization by physicians by themselves.

 

VR Training – this activity helps make the connection between the VR Engine and VR Vocabulary for an individual’s voice.  The more you train the system, the better the quality of the recognition.  The results of this training must be resident locally for the VR Engine doing the processing.  Scribe’s web based interface allows physicians to train locally on their computers and easily transfer this information to our systems for batch processing needs.

 

Voice Recognition Processing Types

Interactive – this is where the physician sits in front of their computer and dictates directly into the Scribe online system.  The editing and approval is done directly by the physician interacting with the online system.

 

Batch – this is where the physician uses Scribe’s VR enabled Digital Dictation Telephony platform.  They make phone calls to the system and are prompted for their dictations.  The information is processed by the Scribe Batch VR system and sent to the physician for editing and review in a short period of time like typical transcriptions.

 

 

Setting Up your Computer for VR

In order to utilize either type of VR with Scribe you’ll need to follow the following instructions:

 

Phase I – Initial VR Release - Interactive

  1. Make sure your computer is running Microsoft XP as the operating system.
  2. If you intend to use VR with Word you’ll need to be running Microsoft Office XP (2002).
  3. You need to enable your computer for speech by following these steps:
    1. Make sure you have either a built-in microphone or an external microphone connected to your computer.
    2. Make sure that the microphone volume is not muted.
    3. Go to the control panel in XP. If you use Category View select the “Sounds, Speech…” category and then the “Speech” icon, if you use the Classic View simply select the “Speech” icon.
    4. A speech properties window will open.  First click on “train microphone” at the bottom to make sure your microphone is set properly.  Follow the prompts and instructions to complete that process.
    5. When you are done configuring your microphone click on the “Train Profile…” button.  Follow the instructions to train your system for VR.  We recommend you do the basic training along with at least 2 additional exercises that will be presented at the end of the initial training.
  4. The language bar which controls the speech recognition should be displayed.  If it is not, Open Regional and Language Options in Control Panel. On the Languages tab, under Text services and input languages, click Details. Under Preferences, click Language Bar.  Select the Show the Language bar on the desktop check box.
  5. The language bar is a floating (you can move it around your screen) tool bar that allows you to start / stop / control voice recognition.  It can also be minimized to the task bar.
  6. There are two types of modes used by VR: Dictation and Voice Commands.  In Dictation mode what you speak is translated to text.  While in dictation mode the computer will also recognize certain commands to help with the editing.  (See appendix A for a complete list). In Voice Command mode the computer will respond to different spoken words to control the particular application you are in.  For Scribe, it is currently only necessary to use the Dictation mode.
  7. You are now ready to begin VR with Scribe. 
    1. Bring up Internet Explorer
    2. Login to the Scribe web site
    3. Select a document to work with or create a new one
    4. Position the cursor where you’d like to begin your dictation
    5. Click the microphone button to “turn on” your microphone
    6. Say “Dictation” if the language bar shows Voice Command mode
    7. Begin your dictation
    8. Click the microphone button on the language bar when you want to pause the dictation.

 

 

Phase II – Batch Processing & Vocabulary Files

In January phase II of the Scribe system will begin.  At that time you will be given further instructions on how to use the Scribe web site to download and use the complete Scribe VR vocabulary files for your personal use and upload your trained profiles to use for batch processing mode.

 

Appendix A - Commands

Basic input commands

Say

To

"New line"

Start text on the next line.

"New paragraph"

Start a new paragraph.

"Tab"

Press the TAB key once.

"Enter"

Press the ENTER key once.

"Spelling mode"

Spell out the next word. For example, say this before you spell out a company name or a person's name. Pause after spelling out the word to revert to normal dictation mode.

"Forcenum"

Enter a number or symbol instead of spelling it out. For example, say "Forcenum 2" to enter "2" instead of "two." Pause after saying the number or symbol to revert to normal dictation mode.

"Space"

Insert a space.

"Backspace"

Delete the character to the left of the insertion point.

"Next cell"

Move the insertion point one cell to the right.

"Spell it"

Spell the next word. If the insertion point is within a word, add spelled letters within the word.

"Spell that"

Spell the next word. If the insertion point is within a word, select the current word and replace it with the spelled word.

"Spelling mode"

Go into or out of spelling mode. In spelling mode, you must spell out the words you want entered as text.

The remaining commands in this topic are supported only by programs that offer complete support for advanced text services, such as Microsoft Word 2002.

Selection commands

Say

To

"Select word" or "Select phrase"

Select the word or phrase given.

"Select word"

Select the last recognized word.

"Select that"

Select the last recognized word or phrase.

"Select word through word"

Select a phrase by specifying the first and last words of the phrase.

"Select previous phrase"

Select the nearest recognized phrase to the left of the insertion point.

"Select next phrase"

Select the nearest recognized phrase to the right of the insertion point.

"Unselect that"

Unselect the currently selected text.

Correction commands

Say

To

"Correct word" or "Correct phrase"

Select the word or phrase given and open the alternate list to supply correction options.

"Correction"

Select the selected text or the last recognized phrase and open the alternate list to supply correction options.

"Correct that"

Select the last recognized word or phrase and open the alternate list to supply correction options.

"Correct previous phrase"

Select the nearest recognized phrase to the left of the insertion point and open the alternate list to supply correction options.

"Correct next phrase"

Select the nearest recognized phrase to the right of the insertion point and open the alternate list to supply correction options.

"Insert after word" or "Insert after phrase"

Move the insertion point to the right of the word or phrase given.

"Insert before word" or "Insert before phrase"

Move the insertion point to the left of the word or phrase given.

"Scratch that"

Delete the last recognized phrase.

"Delete"

Delete the selected word of phrase.

"Delete word"

Delete a specific word.

"Delete word through word"

Delete a phrase by specifying the first and last words of the phrase.

"Reconvert"

Replace the last dictated word or phrase by reprocessing the dictated speech.

Editing commands

Say

To

"Cut"

Delete and copy the selected text to the Clipboard.

"Copy"

Copy the selected text to the Clipboard.

"Paste"

Paste the text into the Clipboard at the insertion point.

Navigation commands

Say

To

"Go to top"

Move the insertion point to the beginning of the page.

"Go to bottom"

Move the insertion point to the end of the page.

"Go to beginning of line"

Move the insertion point to the beginning of the current line.

"Go to end of line"

Move the insertion point to the end of the current line.

"Insert after word" or "Insert after phrase"

Move the insertion point to right of the word or phrase given.

"Insert before word" or "Insert before phrase"

Move the insertion point to left of the word or phrase given.

Keyboard simulation commands

Say

To

"New line"

Start text on the next line.

"New paragraph"

Start a new paragraph.

"Enter"

Press the ENTER key once.

"Tab"

Press the TAB key once.

"Space"

Insert a space at the insertion point.

"Backspace"

Delete the character to the left of the insertion point.

"Next cell"

Move the insertion point one cell to the right.

"Move up"

Move the insertion point up one line.

"Move down"

Move the insertion point down one line.

"Move left"

Move the insertion point one character to the left.

"Move right"

Move the insertion point one character to the right.

"Page up"

Move the insertion point up one page.

"Page down"

Move the insertion point down one page.

Capitalization commands

Say

To

"Capitalize" or "Cap that"

Capitalize the first letter of every word in the last phrase or the selected text.

"Uncapitalize" or "No caps that"

Make all letters lowercase in the last recognized phrase or the selected text.

"All caps that"

Capitalize all letters in the last phrase recognized or the selected phrase.

"Cap it phrase"

Capitalize the first letter of the first word in the phrase.

"All caps phrase"

Capitalize all the letters of the first word in the phrase.

"No caps phrase"

Make all letters in the first word of the phrase lowercase.

Punctuation marks and other symbols

You can say the names of punctuation marks to enter those marks. For example, you can say "comma" to add a comma, and you can say "period" to add a period at the end of a sentence. The following table lists most keyboard characters and how to say them when you are dictating.

To insert

Say

.

"Period" or "dot"

,

"Comma"

:

"Colon"

;

"Semi-colon"

?

"Question mark"

!

"Exclamation point"

&

"Ampersand"

*

"Asterisk"

@

"At sign" or "at"

\

"Backslash"

/

"Slash"

|

"Vertical bar"

-

"Hyphen" or "dash"

--

"Double dash"

=

"Equals"

+

"Plus" or "plus sign"

#

"Pound sign"

%

"Percent" or "percent sign"

$

"Dollar sign"

_

"Underscore"

~

"Tilde"

...

"Ellipsis"

> 

"Greater than"

< 

"Less than"

^

"Caret"

[

"Open bracket"

]

"Right bracket"

{

"Open brace"

}

"Close brace"

(

"Paren"

)

"Close paren"

"

"Quote"

'

"Single quote"

Language bar commands

Say

To

"Microphone"

Turn the microphone off.

"Voice Command"

Switch from Dictation mode to Voice Command mode.

"Correction"

Select the selected text or the last recognized phrase and open the alternate list to supply correction options.