What is speech recognition? Application of speech recognition.

What is Automatic Speech to Text 
Recognition System??? Or what is speech recognition?

Speech Recognition (is also known as Automatic Speech Recognition (ASR), or 

computer speech recognition) is the process of converting a speech signal to a 

sequence of words, by means of an algorithm implemented as a computer 

program. 

Applications of speech recognition 

• In workplaces: Small task commands such as print, record, stop etc. 

• In banking: voiced request, transaction details, making payments etc.

• Search Engines: Voiced searching.

• Language learning

• In internet of Things

• Disabled Persons: Such as typing with speech, running devices with voiced 

commands etc.

Speech Recognition Model

 Speech is the one’s thought and feeling by articulated sounds in the form of 

waves recorded with help of equipment (mic).

 Preprocessing is the cancelation of reverberation, echoing , windowing , noise 

filtering, smoothing and many more.

 Feature extraction, as each and every person has different speech and 

intonation due to different ingrained in the utterances. Reduced the 

variation to a generalized form. Techniques used now a days are LPC and 

MFCC.

 Speech classification Techniques such HMM,DTW, VQ etc.

 Speech recognition : here comes the application of neural networks as these 

were powerful in recognition and there are various NN for this are RNN, LSTM, 

DN and a Hybrid HMM-LSMT. 

Neural Network

Neural Neworks are referred to neuron circuits. At present it is refers to ANN 

consisting of artificial neurons or nodes. Each node receives an input and changes 

it state according to input values and give output. In each node or neuron have a 

mathematical model for information processing.

Generally output of previous neuron is the input for next such a type of NN is 

called as FFNN(feed forward neural network) and having a DAG(directed Acyclic 

Graph).

Networks having cycle in it called as RNN

Conclusion

Neural networks are modern way for recognizing speech. SRS requires four stages

Analysis, feature extraction , modeling and matching. ASR using neural network

is emerging field now a day. Various neural networks such as deep neural

network, RNN and LSTM are used for recognizing speech.