Principal Feature Extraction from Regional Speech in Emergency Communication
Mostafizur Rahaman Laskar, Prasanta Kr Sen, Ishita Paul, Saptargha Das, Sudipta Kumar Ghosh · 2019 Second International Conference on Advanced Computational and Communication Paradigms (ICACCP) · 2019
We have presented a method of extracting principal features from the speech signal of a few key emergency words frequently used in Eastern India. Based on the collected sample of different emergency situations such as fire, flood, road accident, medical emergencies(eg. cardiac stroke, unconsciousness) from multiple subjects (Persons) from Eastern India (West Bengal, Odisha, and Assam). Principal Component Analysis (PCA) based feature extraction method is used to extract the features from the samples. The statistical parameters (Mean, Variance) of the principal feature components are estimated using the maximum likelihood estimate(MLE) of the speech signal. The proposed PCA based algorithm may help to recognize speech in emergency communication.