Virtual Personal Assistant
Designed by : Siddhesh Ravindra Nanche
Introduction
• A v irt ua l a ssist a nt is a so f tw ar e ag en t th at can
p er f o rm task s o r ser v ices f o r an in d iv id ual.
• To activ ate a v ir tu al assistan t u sin g th e v o ice, a w ak e
w o r d mig h t b e u sed .
• Vir tu al assistan ts can p r o v id e a w id e v ar iety o f
ser v ices.
• I t main ly u ses Sp eech r eco g n ition tech n iq u e.
Basic Structure
of the System
Algorithms Used
• Training Algorithm
• Pattern Recognition Algorithm
• Voice Recognition Algorithms
• Word Selection Algorithm
Processing of speech signals
Feature Extraction Methods
Perceptual Linear Prediction(PLP)
Relative spectra filtering of log domain coefficient PLP(RASTA-PLP)
Linear Predictive Coding(LPC)
Predictive cepstral coefficient(LPCC)
Mel scale cepstral Analysis(MEL)
Mel frequency cepstral coefficient(MFCC)
Method Resources
required
Popularity Implementa
tion
Multi
Speaker
Multi
Language
PLP LOW HIGH EASY NO NO
LPCC HIGH LOW DIFFICULT YES YES
MFCC MODERATE HIGH MODERATE YES YES
Comparison of Feature Extraction Methods
Voice Recognition
Algorithms using
M e l
F r e q u e n c y
C e p s t r a l
C o e f f i c i e n t
(MFCC) method
Start
Record Voice Command
Get MFCC vector & Store into
reference template
Get measuring similarity between training & testing
input voice signal
Received voice command
Match with
reference
template
Send the signal to activated command
No
End
yes
Neural Network
General Diagram of neural network
structure of a neural network with a feedback.
Equations used in calculation of network
F o r H i d d e n L ay e r
F o r N o n - L i n e a r A c t i v a t i o n F u n c t i o n
F o r E r ro r S y s t e m F u n c t i o n
ADVANTAGES
Work processes become more efficient
 learns as it works
improved even further
Who engage in mobile dictation need scarcely change their routine
improves efficiency, leads to more structured work
Applications
• Google Now
• Google Assistant
• Apple’s Siri
• S Voice
• Microsoft’s Cortana
• Mycroft AI
Conclusion
 Virtual Personal Assistant has very importance now-a-days. Use of this leads to
smart work.
 It gives efficient way to communicate with hand-held device.
 There are many ways to use or implement this system.
 This system learns as it works, and improves efficiency.
Thank You

Virtual personal assistant

  • 1.
    Virtual Personal Assistant Designedby : Siddhesh Ravindra Nanche
  • 2.
    Introduction • A virt ua l a ssist a nt is a so f tw ar e ag en t th at can p er f o rm task s o r ser v ices f o r an in d iv id ual. • To activ ate a v ir tu al assistan t u sin g th e v o ice, a w ak e w o r d mig h t b e u sed . • Vir tu al assistan ts can p r o v id e a w id e v ar iety o f ser v ices. • I t main ly u ses Sp eech r eco g n ition tech n iq u e.
  • 3.
  • 4.
    Algorithms Used • TrainingAlgorithm • Pattern Recognition Algorithm • Voice Recognition Algorithms • Word Selection Algorithm
  • 5.
  • 6.
    Feature Extraction Methods PerceptualLinear Prediction(PLP) Relative spectra filtering of log domain coefficient PLP(RASTA-PLP) Linear Predictive Coding(LPC) Predictive cepstral coefficient(LPCC) Mel scale cepstral Analysis(MEL) Mel frequency cepstral coefficient(MFCC)
  • 7.
    Method Resources required Popularity Implementa tion Multi Speaker Multi Language PLPLOW HIGH EASY NO NO LPCC HIGH LOW DIFFICULT YES YES MFCC MODERATE HIGH MODERATE YES YES Comparison of Feature Extraction Methods
  • 8.
    Voice Recognition Algorithms using Me l F r e q u e n c y C e p s t r a l C o e f f i c i e n t (MFCC) method Start Record Voice Command Get MFCC vector & Store into reference template Get measuring similarity between training & testing input voice signal Received voice command Match with reference template Send the signal to activated command No End yes
  • 9.
  • 10.
    General Diagram ofneural network
  • 11.
    structure of aneural network with a feedback.
  • 12.
    Equations used incalculation of network F o r H i d d e n L ay e r F o r N o n - L i n e a r A c t i v a t i o n F u n c t i o n F o r E r ro r S y s t e m F u n c t i o n
  • 13.
    ADVANTAGES Work processes becomemore efficient  learns as it works improved even further Who engage in mobile dictation need scarcely change their routine improves efficiency, leads to more structured work
  • 14.
    Applications • Google Now •Google Assistant • Apple’s Siri • S Voice • Microsoft’s Cortana • Mycroft AI
  • 15.
    Conclusion  Virtual PersonalAssistant has very importance now-a-days. Use of this leads to smart work.  It gives efficient way to communicate with hand-held device.  There are many ways to use or implement this system.  This system learns as it works, and improves efficiency.
  • 16.