× Close
Join the Ques10 Community
Ques10 is a community of thousands of students, teachers, and academic experts, just like you.
Join them; it only takes a minute
Sign up
Question: Write the state of art of speech recognition
0

Subject: Speech Processing

Topic: Speech Processing Applications

Difficulty: Low

sp(38) • 359 views
ADD COMMENTlink
modified 5 months ago by gravatar for Manan Bothra Manan Bothra0 written 6 months ago by gravatar for awari.swati831 awari.swati831140
0

(i) Tremendous advances have been observed in ASR systems in the past decades. The word recognition error rates has reduced by factor of 5.

(ii) Fast recognition algorithm have been developed into increase the recognition rate several time.

(iii) Speaker-independent speech recognition is made possible with high accuracy without extra hardware.

(iv) Continuous speech-recognition is also advanced in the following respects:

a) Speech modelling techniques have been improved to give higher speech recognition accuracy.

b) Search strategies have been developed to reduced the recognition time.

(v) These advances have occurred not only in laboratories, but speech recognition is strongly influencing the communication between human and machines.

(vi) Recent speech recognition uses probabilistic modelling methods, such as Hidden Markov Modelling (HMM).

(vii) These methods can also model the time variation of the spectral characteristics of speech signals. The model parameters can be found automatically from the training data. The probabilistic process has optimized the recognition accuracy.

(viii) The recognition error rates are found to reduce with the probabilistic modelling.

(ix) It is required to continue the research to find better models that are robust with respect to speaker variability, noise acoustic distortions.

ADD COMMENTlink
written 5 months ago by gravatar for Manan Bothra Manan Bothra0
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.