Voice Recognition Techniques for Front End in Vb.Net
Voice recognition and speech recognition occur when a computer translates human speech into text or accepts them as commands. Speech recognition permits hands free usage of a computer. Virtual Basic.net, or vb.net for short, is a software language developed by Microsoft and is used for voice and speech recognition software. Front end VB voice recognition accepts speech as its input.
Other People Are Reading
-
Uses
-
Speech recognition software acts as a front end filter to handle as many calls as possible before forwarding customers to the call center. Speech recognition is used for digital dictation. Voice recognition can convert words into text while translating the text into another language, providing real-time translation. Voice recognition is used in automated phone systems that respond to callers’ requests for information. Voice recognition accepts commands from users who cannot enter commands via a keyboard.
Existing Utilities
-
Microsoft Word contains a built-in speech utility for dictating the user’s speech to text. “Succeeding with Technology” states that Microsoft Windows and Mac OS X have utilities that allow the user to speak commands to the computer. According to “Windows Speech Recognition Programming” by Keith Jones, Microsoft Speech API standardizes the "sound-and-voice" of Windows speech recognition programming. The Microsoft speech constants are stored in the vbspeech.bas file.
-
Windows Presentation Foundation
-
Speech synthesis and speech recognition in Windows Presentation Foundation
(WPF) are part of the Microsoft vb.net framework 3.0. Microsoft speech automation services are performed using visual basic activeX. According to “Pro WPF in VB 2010” by Matthew MacDonald, “The most straightforward way to use speech recognition is to create an instance of the SpeechRecognizer class from the System.Speech.Recognition namespace.”
Voice Recognition Logic
-
Voice recognition software must decode each set of sounds and translate them to the correct word. Natural language processing applies artificial intelligence logic to decoding human speech. According to “Technology Guide: Principles, Applications, Trends” by Hans-Jörg Bullinger, “A central method in the recognition of continuously spoken phrases is the use of so-called Hidden-Markov-Models (HHM) – stochastic models used to describe and analyze real signal sources and to extract the phonemes that fit the input signal best.” Artificial intelligence in voice recognition can be trained. As the voice recognition software records the user’s words and is corrected when it incorrectly identifies a word, its accuracy improves.
Types of Spoken Commands
-
Command control dictation uses a limited set of spoken commands. Speech and voice recognition uses commands in the natural language of a user. Command control dictation would use a command like “Open Microsoft Word, use 12 point italic font.” Voice recognition software intelligent enough to accept natural language commands could follow the command, “Open MS Word and switch to 12 point italic font.”
Preventing Problems
-
Eliminate background noise to reduce the errors created by sound picked up as speech or interfering with voice recognition. Speaker-independent voice recognition software does not require a training period. However, frequently changing users of speaker-dependent voice recognition software increases the risk of errors. Voice recognition errors are decreased with increased training time with each individual user.
-
References
Resources
- Photo Credit Jack Hollingsworth/Photodisc/Getty Images Comstock/Comstock/Getty Images
